Machine Learning | Techlist.io

grammarlyFeb 13, 2026Original article

What Is an AI Assistant? Definition, Types, and Examples (opens in new tab)

AI assistants have evolved from simple command-driven tools into sophisticated digital partners that leverage natural language processing to streamline workplace productivity. By integrating large language models with real-time data and contextual awareness, these tools enable users to automate repetitive tasks and manage information more effectively. Ultimately, their value lies in their ability to bridge the gap between open-ended human intent and actionable digital output across diverse software environments. ### The Technical Framework of AI Interaction * **Natural Language Processing (NLP):** This technology allows assistants to interpret the nuance of everyday language, distinguishing between literal questions and requests for tonal adjustments or stylistic changes. * **Large Language Models (LLMs):** These models use machine learning patterns to predict and generate helpful responses rather than relying on a pre-written script. * **Context Windows:** Modern assistants maintain a "memory" of the current conversation or document, allowing them to refer back to earlier sections and maintain consistency across long-form projects. * **Tool Integration:** Many assistants function by connecting to external APIs, enabling them to check calendars, pull data from the web, or manage task lists within other applications. ### Functional Applications in Daily Workflows * **Content Synthesis:** Assistants can ingest lengthy documents or meeting recordings to produce condensed summaries, outlines, and key takeaways. * **Drafting and Revision:** Beyond simple generation, these tools help refine existing text for clarity, length, and professional tone. * **Ideation and Brainstorming:** Users can utilize AI to overcome the "blank page" problem by generating initial project structures or exploring different angles for a specific topic. * **Technical Support:** For developers, AI assistants can interpret error messages, generate code snippets, and explain complex technical concepts in plain language. To maximize the impact of these tools, users should focus on providing detailed prompts that provide clear context and intent. As AI assistants become more deeply embedded in browsers and operating systems, understanding the balance between their generative capabilities and their contextual limitations is essential for maintaining an efficient digital workflow.

Machine Learning Large Language Models Natural Language Processing Productivity Tools+3

grammarlyFeb 13, 2026Original article

How to Create an AI Assistant Step by Step: A Beginner’s Guide (opens in new tab)

Creating a custom AI assistant is no longer restricted to engineers, as modern no-code tools and APIs allow users to build specialized agents for specific personal or professional workflows. By focusing on a narrow scope and selecting the right platform, individuals can gain greater control over data, behavior, and task efficiency than generic tools provide. Ultimately, the shift toward custom assistants reflects a move away from one-size-fits-all software toward personalized AI teammates integrated directly into daily work. ## The Anatomy of an AI Assistant * Digital assistants utilize Natural Language Processing (NLP) to interpret user intent and tone through conversational prompts. * Large Language Models (LLMs) serve as the underlying engine, recognizing language patterns to generate contextually relevant responses. * Advanced implementations, such as the "Go" assistant, operate within existing apps like email and documents to eliminate context switching and manual data entry. ## Strategic Drivers for Customization * **Personalization:** Tailoring the assistant’s tone and behavior ensures it supports specific tasks exactly as the user expects. * **Data Control:** Building a custom solution offers transparency into how data is used, which is critical for teams handling sensitive internal information. * **Efficiency and Innovation:** Customizing an assistant for a niche problem—like summarizing specific document types or automating recurring questions—reduces manual effort more effectively than general tools. * **Independence:** Creating a proprietary tool reduces reliance on third-party platforms that may change their pricing or feature sets. ## Defining the Core Mission * The most successful assistants focus on one primary responsibility rather than trying to handle every possible task. * Effective planning requires answering who the user is and what specific problem the assistant is meant to solve consistently. * Starting with a narrow scope, such as a dedicated writing assistant or a customer service bot, simplifies the testing and refinement process during the initial launch. ## Development Paths and Lifecycles * Users can choose between no-code platforms for rapid deployment or API-based configurations for higher flexibility and integration. * The development process follows a standard lifecycle: strategic planning, technical configuration, launch, and continuous improvement. * Ongoing monitoring is essential to ensure the assistant remains responsible, accurate, and aligned with evolving user needs. To build a successful AI assistant, start by identifying a single high-impact task and selecting a tool that matches your technical comfort level. Prioritizing a narrow focus during the initial build will allow for more effective monitoring and easier scaling as your requirements grow.

Machine Learning Large Language Models Automation Natural Language Processing+3

googleFeb 9, 20263 min readCurated summary

How AI trained on birds is surfacing underwater mysteries

Perch 2.0, Google DeepMind’s bioacoustics foundation model, was trained mainly on birds and terrestrial animals yet performs strongly on underwater audio. The study shows that its learned audio embeddings can support accurate whale, dolphin, reef-sound, and killer-whale classification with only a few labeled examples. This suggests that large, broadly trained bioacoustics models can transfer across environments and accelerate marine research without requiring extensive underwater training data. ## Underwater Mysteries and Bioacoustics - Ocean recordings reveal animal behavior, species distributions, and unexplained sounds. - The “biotwang,” recently attributed by NOAA to Bryde’s whales, illustrates how new calls and species identifications continue to emerge. - Google has previously developed models for humpback whales and multi-species whale detection. - Perch 2.0 extends this work despite having no underwater audio in its training data. ## Transfer Learning for Custom Classifiers - Researchers can use an existing model directly when its labels match their data. - For new sounds or datasets, transfer learning avoids training a deep neural network from scratch. - Perch 2.0 converts audio windows into compact numerical embeddings. - A logistic regression classifier is then trained on those embeddings using labeled examples. - This requires far less computation, experimentation, and training data than full neural-network training. ## Evaluation on Marine Datasets - The researchers tested Perch 2.0 with few-shot linear probes using 4, 8, 16, or 32 examples per class. - Performance was measured using ROC-AUC, where values closer to 1 indicate better class separation. - Evaluation datasets included: - **NOAA PIPAN:** Baleen-whale recordings, including minke, humpback, sei, blue, fin, and Bryde’s whales. - **ReefSet:** Reef biological sounds, fish, dolphins, anthropogenic noise, and waves. - **DCLDE:** Killer whales, humpbacks, abiotic sounds, unknown sounds, and killer-whale ecotypes. - More examples generally improved results. - ReefSet performance was already high with four examples per class for most models. - Perch 2.0 was consistently among the best-performing models across datasets and sample sizes. ## Comparisons with Other Models - Perch 2.0 was compared with Perch 1.0, SurfPerch, and Google’s multi-species whale model. - It also outperformed AVES-bird and AVES-bio on most underwater tasks. - The results show that strong underwater transfer is not limited to models trained on marine audio. ## Why Bird-Based Training Transfers to Whales - The authors suggest that large models trained on extensive datasets can generalize effectively to unfamiliar downstream tasks. - Shared acoustic patterns across animal vocalizations may allow representations learned from birds and other terrestrial species to remain useful underwater. - The findings challenge the assumption that a model must be trained directly on underwater recordings to perform well on marine classification tasks. ## Practical Tools for Researchers - Google provides a paper and a Google Colab tutorial. - The tutorial demonstrates an end-to-end workflow for building a whale-vocalization classifier. - It uses NOAA’s NCEI Passive Acoustic Data Archive and Google Cloud. - Researchers can create agile, task-specific models with relatively small labeled datasets. Perch 2.0 demonstrates that broad bioacoustic pretraining can substantially reduce the effort required to study marine sounds. Researchers can begin with general-purpose embeddings and adapt them to new whale species, calls, or underwater sound categories using only modest labeled data.

Read original

Machine Learning Foundation Models Embeddings Few Shot Learning+4

spotifyFeb 9, 20263 min readCurated summary

Our Multi-Agent Architecture for Smarter Advertising | Spotify Engineering

The post argues that fragmented advertising workflows, not backend infrastructure, are the core problem. Although buying channels share services and data, their planning and optimization logic is repeatedly reimplemented across channels and surfaces, causing drift and technical debt. The proposed solution is a shared agentic decision layer that interprets advertiser goals, orchestrates existing Ads APIs, and applies consistent reasoning across products. ## Fragmented Workflows Across a Shared Backend - Direct, Self-Serve, and Programmatic buying use largely consolidated infrastructure but retain different workflows and decision logic. - Spotify Ads Manager, Salesforce, Slack, and internal tools contain overlapping automation. - Budget allocation, inventory selection, reach, efficiency, and STR decisions are repeatedly implemented in different places. - Incremental workflow changes therefore create duplicated maintenance work and inconsistent behavior. ## Why Conventional Workflow Services Fall Short - Hard-coded state machines and REST services are poorly suited to combinatorial planning tasks. - Campaign planning depends on: - User and advertiser characteristics - Available inventory and audiences - Business priorities - Forecasts, performance, and optimization goals - A workflow optimized for one channel or “happy path” will not adapt well as requirements change. - Improvements to decision logic must be replicated across every product surface, increasing the risk of divergence. ## The Missing Intent Layer - Existing systems can perform individual actions such as creating line items, running forecasts, and retrieving insights. - They do not consistently translate high-level objectives into: - A sequence of tool calls - Explicit tradeoffs - Validation and safety checks - An objective such as maximizing reach in Brazil while protecting video inventory and meeting STR requires coordinated reasoning across multiple capabilities. ## A Modular Agentic Architecture - Campaign planning and management are modeled as cooperating specialized agents. - Agents use shared signals, including: - Inventory - Audiences - STR - Quality and risk - Historical performance - They jointly optimize advertiser goals and Spotify’s business constraints. - Existing Ads services become tools that agents orchestrate, rather than capabilities being rebuilt in each workflow. - A long-running orchestration layer delegates tasks while agents share context and evaluation logic. - The same decision engine can support every buying channel and surface. ## Engineering Implications - APIs need to be designed as agent tools, rather than only as CRUD interfaces. - Testing must include behavioral evaluation in addition to unit and integration tests. - Observability should explain what an agent decided and why, not merely track latency and errors. - Safety requires guardrails for semi-autonomous decisions, beyond ordinary input validation. - The approach avoids both duplicated deterministic workflows and a brittle, centralized rules engine for probabilistic, ML-heavy advertising logic. The overall recommendation is to centralize campaign decision-making in a reusable agentic platform while keeping existing services as specialized tools. This should reduce duplicated workflow logic, make improvements consistent across products, and allow advertising workflows to evolve without repeatedly rebuilding them.

Read original

Machine Learning AI Agents API Design Multi Agent Systems+3

pinterestFeb 2, 20264 min readCurated summary

Beyond Two Towers: Re-architecting the Serving Stack for Next-Gen Ads Lightweight Ranking Models…

Two-Tower models make retrieval and lightweight ranking highly efficient by scoring user and item embeddings with a dot product, but they cannot represent rich user-item interactions or deep feature crossings. This post describes an ads-serving redesign that introduces general-purpose GPU models while preserving end-to-end latency. The main strategy is to reduce data movement, move filtering logic onto the GPU, and optimize inference from an initial 4-second p90 latency to about 20 milliseconds. ## Why Move Beyond Two-Tower Models - Two-Tower architectures independently encode users and items, enabling fast scoring across millions of candidates. - Their decoupled structure limits: - User-item interaction features - Target attention - Early feature crossing - Deep architectures requiring simultaneous access to user and candidate data - More expressive models require GPU-based general-purpose inference rather than specialized dot-product or ANN retrieval. - The existing retrieval stack was not designed to transfer large candidate and feature sets to a GPU, creating a major latency challenge. ## Restructuring the Serving Funnel The traditional funnel consisted of: - Feature expansion for thousands of candidates - Retrieval and Two-Tower lightweight ranking - Heavy ranking and auction processing for the top documents Adding GPU inference directly to this flow would require fetching, serializing, transferring, and returning features for tens of thousands of documents. The authors therefore redesigned the entire early-stage serving pipeline instead of optimizing the model alone. ## Segmenting the Inventory for Feature Fetching Feature retrieval was a major latency source, often taking longer than model inference for workloads ranging from 10,000 to 100,000 documents. - **High-value inventory:** Roughly 1 million documents responsible for a substantial share of revenue have their features embedded in the PyTorch model as registered buffers. - Features become part of the model state, similar to weights. - They remain in GPU high-bandwidth memory. - Requests avoid remote feature-service calls and host-to-device transfers. - The model file must be periodically updated to refresh features. - Future work may include GPU-based caching. - **Long-tail inventory:** The remaining roughly 1 billion documents use a high-performance key-value store with in-host caching. - The post focuses on the first strategy, which is already running in production. ## Moving Business Logic onto the GPU Previously, the model returned scores for approximately 100,000 candidates, while CPU-side code handled utility calculation, filtering, diversity, deduplication, and top-k selection. - The new PyTorch model performs these operations directly: - Combines pCTR, pCVR, bid, and other signals into utility scores. - Applies diversity and filtering rules. - Performs top-k selection. - The GPU returns only the final winners—typically around 1,000 documents—instead of all candidate scores. - This reduces device-to-host data transfer and takes advantage of GPU parallelism. - The approach works because lightweight-ranking business rules are sufficiently simple to express with tensor operations. ## Reducing GPU Inference Latency Initial GPU inference measured roughly 4,000 ms at p90, far too slow for real-time serving. Several systems optimizations reduced this to approximately 20 ms: - **Multiple CUDA streams:** Separate streams for workers allow host-to-device transfers, computation, and device-to-host transfers to overlap. - **Worker alignment:** Worker threads are matched and pinned to physical CPU cores to reduce context switching and lock contention. - **Kernel fusion:** Triton kernels combine operations such as linear layers and activations, reducing memory traffic. - **BF16 computation:** Brain Floating Point 16 lowers memory usage and accelerates arithmetic compared with FP32. - **Profiling tools:** PyTorch Profiler and NVIDIA Nsight Systems were used to identify bottlenecks. ## Practical Recommendation Deploying more expressive ranking models requires rethinking the serving architecture around data movement and execution placement. Embedding frequently used features, executing business logic on the GPU, and applying low-level CUDA and kernel optimizations can make complex neural ranking feasible without increasing end-to-end latency.

Read original

Machine Learning Recommendation Systems Pytorch Gpu Inference+4

googleJan 23, 20263 min readCurated summary

Introducing GIST: The Next Stage in Smart Sampling | Google Research

GIST is a data-subset selection algorithm designed to balance diversity and utility when training on massive datasets. It converts the difficult diversity–utility optimization problem into a series of thresholded graph problems and uses a bicriteria greedy strategy to find a high-quality subset efficiently. The algorithm guarantees at least half the value of the optimal solution, while the authors prove that improving beyond a 0.56 approximation is NP-hard. ## Why Smart Sampling Is Difficult - Large ML systems need to process datasets that are increasingly expensive to store, analyze, and train on. - Subset selection aims to choose a smaller but representative set of examples. - **Diversity** prevents redundant selections by maximizing the minimum distance between selected points, typically in embedding space. - **Utility** measures how much relevant or unique information the subset provides, modeled using monotone submodular functions. - Optimizing both objectives simultaneously is NP-hard: - A diversity-only method may select irrelevant examples. - A utility-only method may select many similar examples from one highly relevant cluster. ## How GIST Works ### Diversity Thresholding - GIST fixes a candidate minimum distance rather than optimizing the distance directly. - It builds a graph in which two data points are connected when they are closer than the chosen threshold. - Connected points are considered too similar to coexist in the selected subset. - Selecting points that are not connected enforces the desired spacing between examples. ### Utility-Constrained Independent Sets - For each threshold, GIST seeks a high-utility independent set: a group of points with no edges between them. - This corresponds to selecting valuable examples without choosing mutually conflicting or overly similar points. - Because maximum independent set is NP-complete and lacks practical general-purpose approximation algorithms, GIST uses a specially designed bicriteria greedy method. - The algorithm repeatedly selects high-scoring points and excludes nearby candidates, effectively creating “no-go zones” around selected data. ### Searching Across Thresholds - GIST evaluates all relevant distance thresholds derived from the dataset. - It greedily constructs a candidate subset for each threshold. - It returns the best candidate found across these runs. - If the optimal solution achieves minimum distance $d$, GIST obtains comparable utility while guaranteeing a minimum distance of roughly $d/2$. ## Theoretical Guarantees - GIST is presented as the first algorithm with a strong provable guarantee for this diversity–utility tradeoff. - Its output has at least half the value of the absolute optimum. - The authors also prove that finding a solution worth more than 0.56 of the optimum is NP-hard. - These results provide a mathematical guarantee that GIST is not merely producing empirically good subsets, but making a bounded tradeoff between informativeness and coverage. ## Practical Evaluation - GIST was evaluated against several common subset-selection approaches in ML applications. - Comparisons included: - **Random**, a simple baseline that often provides reasonable diversity. - **Margin**, which selects examples the model is uncertain about but does not explicitly promote diversity. - **k-center**, which minimizes representation gaps by keeping all data points close to a selected representative. - **Submod**, which combines utility with an older formulation of diversity. - The experiments, including image-classification benchmarks, reportedly show that GIST outperforms state-of-the-art alternatives while retaining formal guarantees. GIST is therefore a practical choice when subset selection must preserve both broad data coverage and task relevance. Its main advantage is combining competitive real-world performance with a clear approximation guarantee, rather than relying solely on heuristic results.

Read original

Machine Learning Approximation Algorithms Subset Selection Data Sampling+4

gitlabJan 15, 2026Original article

Introducing GitLab Credits (opens in new tab)

GitLab is transitioning from seat-based pricing to a usage-based model with the introduction of GitLab Credits, a virtual currency designed for the GitLab Duo Agent Platform. This shift addresses the limitations of traditional licensing, which often creates "AI haves and have-nots" by making access too expensive for light or occasional users. By pooling resources across an entire organization, GitLab aims to provide equitable access to agentic AI for every developer while ensuring costs align with actual consumption. ## The Shift from Seat-Based to Usage-Based AI * Traditional seat-based models are poorly suited for agentic AI, which can be triggered by background SDLC events rather than just direct user interaction. * The credit model allows every member of a Premium or Ultimate organization to use AI capabilities without requiring an individual "AI seat." * Usage-based pricing automatically offsets the costs of power users against lighter users, lowering the total cost of ownership for the organization. ## Mechanics of GitLab Credits * Credits function as a pooled resource consumed by both synchronous interactions (like Agentic Chat in the IDE) and asynchronous background tasks. * Supported capabilities include foundational agents (Security, Planner, Data Analyst) and specific workflows such as Code Review and CI/CD pipeline fixing. * The system integrates with external models like Anthropic Claude Code and OpenAI Codex, as well as custom agents published in the GitLab AI Catalog. * Each credit has an on-demand list price of $1, with volume discounts available for enterprise customers who sign up for annual commitments. ## Governance and Usage Controls * Administrators can monitor consumption through two dedicated dashboards: a financial oversight portal for billing managers and an operational monitoring view for administrators. * Granular controls allow organizations to enable or disable Duo Agent Platform access for specific teams or projects to prevent unexpected credit depletion. * Proactive email alerts are triggered when consumption reaches 50%, 80%, and 100% of committed monthly credits. * A sizing calculator is available to help organizations estimate their monthly credit requirements based on patterns observed during the platform's beta period. ## Transitioning and Promotional Access * Existing GitLab Duo Pro and Duo Enterprise customers can roll over their current seat investments into GitLab Credits with volume-based discounts. * As part of a limited-time promotion, GitLab is providing $12 in monthly credits per user for Premium subscribers and $24 per user for Ultimate subscribers. * Self-managed and GitLab Dedicated customers will gain access to these credit-based features starting with the 18.8 and 18.9 releases. For organizations looking to scale AI across the software development lifecycle, the credit-based model offers a more flexible and cost-effective path than rigid seat licenses. Current Premium and Ultimate subscribers should leverage their monthly promotional credits to baseline their usage before committing to larger annual credit bundles.

Machine Learning AI Agents Gitlab Duo Gitlab Credits+4

metaJan 14, 2026Original article

Adapting the Facebook Reels RecSys AI Model Based on User Feedback (opens in new tab)

Meta has enhanced the Facebook Reels recommendation engine by shifting focus from traditional engagement signals, like watch time and likes, to direct user feedback. By implementing the User True Interest Survey (UTIS) model, the system now prioritizes content that aligns with genuine user preferences rather than just short-term interactions. This shift has resulted in significant improvements in recommendation relevance, high-quality content delivery, and long-term user retention. **Limitations of Engagement-Based Metrics** * Traditional signals like "likes" and "watch time" are often noisy and may not reflect a user’s actual long-term interests. * Models optimized solely for engagement tend to favor short-term value over the long-term utility of the product. * Internal research found that previous heuristic-based interest models only achieved 48.3% precision in identifying what users truly care about. * Effective interest matching requires understanding nuanced factors such as production style, mood, audio, and motivation, which implicit signals often miss. **The User True Interest Survey (UTIS) Model** * Meta collects direct feedback via randomized, single-question surveys asking users to rate video interest on a 1–5 scale. * The raw survey data is binarized to denoise responses and weighted to correct for sampling and nonresponse bias. * The UTIS model functions as a lightweight "alignment model layer" built on top of the main multi-task ranking system. * The architecture uses existing model predictions as input features, supplemented by engineered features that capture content attributes and user behavior. **Integration into the Ranking Funnel** * **Late Stage Ranking (LSR):** The UTIS score is used as an additional input feature in the final value formula, allowing the system to boost high-interest videos and demote low-interest ones. * **Early Stage Ranking (Retrieval):** The model aggregates survey data to reconstruct user interest profiles, helping the system source more relevant candidates during the initial retrieval phase. * **Knowledge Distillation:** Large sequence-based retrieval models are aligned using UTIS predictions as labels through distillation objectives. **Performance and Impact** * The deployment of UTIS has led to a measurable increase in the delivery of niche, high-quality content. * Generic, popularity-based recommendations that often lack depth have been reduced. * Meta observed robust improvements across core metrics, including higher follow rates, more shares, and increased user retention. * The system now offers better interpretability, allowing engineers to understand which specific factors contribute to a user’s sense of "interest match." To continue improving the Reels ecosystem, Meta is focusing on doubling down on personalization by tackling challenges related to sparse data and sampling bias while exploring more advanced AI architectures to further diversify recommendations.

Machine Learning AI Recommendation Systems Information Retrieval+5

gitlabJan 14, 2026Original article

Understanding agents: Foundational, custom, and external (opens in new tab)

The GitLab Duo Agent Platform provides a tiered framework for integrating AI into the software development lifecycle through foundational, custom, and external agents. By combining built-in expertise with the ability to define bespoke behaviors or connect to specialized external models, the platform enables teams to automate complex tasks ranging from product planning to runtime debugging. This structured approach ensures that AI assistance is deeply integrated into GitLab’s ecosystem while remaining flexible enough to meet specific organizational standards. ## Foundational Agents These are pre-configured, GitLab-maintained agents available immediately in the IDE or Web UI for general and specialized SDLC tasks. * **GitLab Duo:** The primary general-purpose partner for code modification, merge request management, and issue triaging within the full platform context. * **Planner Agent:** Specifically designed to assist with product management by breaking down epics into structured issues and generating acceptance criteria. * **Security Analyst Agent:** Focuses on triaging vulnerabilities, identifying false positives from scans, and prioritizing risks based on actual impact. * **Data Analyst Agent:** Leverages GitLab Query Language (GLQL) to visualize platform data, such as merge request trends, team workloads, and issue resolution times. ## Custom Agents Organizations can create specialized agents tailored to internal workflows by defining unique system prompts and visibility settings. * **Configuration and Control:** Custom agents are defined by a system prompt that dictates their persona and expertise—such as a DevOps agent that correlates static code data with CI/CD logs. * **Visibility Tiers:** Agents can be set to "Private" for use within a specific project or "Public" to be listed in the AI Catalog for broader organizational discovery. * **Operational Use Cases:** Common implementations include onboarding assistants for company-specific practices, compliance monitors for regulatory requirements, and localized support agents for non-English languages. * **Deployment Best Practices:** It is recommended to start with read-only permissions and highly specific constraints before granting agents write access to the repository or platform. ## External Agents External agents operate asynchronously and are triggered by mentions or assignments within issues and merge requests, rather than through interactive chat. * **Asynchronous Automation:** These agents, such as Anthropic Claude or OpenAI Codex, execute tasks in the background when triggered by commands like `@ai-codex`. * **Managed Credentials:** GitLab handles API key management and rotation for these integrations, simplifying the security overhead for teams using third-party models. * **Specialized Performance:** External agents allow teams to leverage provider-specific strengths, such as Claude’s code analysis or Codex’s task delegation, while maintaining compliance with specific data residency requirements. * **Integrated Review:** A typical workflow involves assigning an external agent as a reviewer on a merge request, where it automatically analyzes code quality and posts improvement suggestions directly as comments. To maximize the value of the platform, teams should begin by leveraging foundational agents for immediate productivity gains before developing custom agents that encode specific organizational knowledge. External agents should be reserved for specialized automation tasks or when specific third-party large language models (LLMs) are required for compliance or advanced code generation.

Machine Learning Database Design CI/CD Gitlab Duo+4

gitlabJan 14, 2026Original article

How to customize GitLab Duo Agent Platform (opens in new tab)

The GitLab Duo Agent Platform provides a multi-layered framework for customizing AI behavior to align with specific team workflows and coding standards. By leveraging configuration files at the user, workspace, and project levels, teams can ensure that AI-driven assistance remains context-aware and adheres to internal development policies. This extensibility allows organizations to move from generic AI interactions to highly specialized automation that respects unique architectural patterns and security requirements. ### Levels of Customization GitLab offers a hierarchical approach to tailoring agent behavior, ensuring the right balance between global consistency and project-specific flexibility: * **User-level:** Personal preferences and rules applied across all projects, typically stored in the user’s home directory (e.g., `~/.gitlab/duo/`). * **Workspace-level:** Project-specific configurations located in the repository root that override user-level settings for that specific codebase. * **Project-level:** The creation of entirely custom agents and workflows managed within a specific project to handle complex, specialized tasks. ### Custom Rule Configuration Custom rules provide a mechanism to enforce specific coding styles and instructional sets without repeating prompts in every interaction. * **File implementation:** Rules are defined in `chat-rules.md` files located either in the user's home directory for global application or within the `.gitlab/duo/` directory for project-specific application. * **Functional scope:** They are best used for granular instructions such as forcing the use of the Vue 3 Composition API, requiring JSDoc comments for public functions, or mandating single quotes for strings. * **Governance:** Teams are encouraged to use GitLab Code Owners to manage who can approve changes to these rules, ensuring that AI behavior remains aligned with official team standards. ### Architectural Control with AGENTS.md The platform supports `AGENTS.md`, an industry-standard configuration file used to define broader agent personality, tone, and deep repository context. * **Versatility:** Unlike basic rules, `AGENTS.md` is consumed by both foundational and custom flows and can be understood by external agents like Claude Code. * **Contextual Depth:** These files can be placed in subdirectories to provide specific instructions for different parts of a monorepo, helping the agent understand complex folder structures and internal dependencies. * **Key Parameters:** It typically controls high-level preferences such as security protocols (e.g., "never suggest hardcoding secrets"), documentation requirements, and preferred tool usage. ### Technical Requirements and Deployment Implementing these customizations requires specific environment versions to ensure compatibility across the GitLab ecosystem. * **GitLab Version:** Requires GitLab 18.8 or later. * **IDE Support:** For VS Code users, the GitLab Workflow extension must be version 6.60 or later. * **Update Cycle:** Changes to `AGENTS.md` or custom rules generally require starting a new chat session or triggering a new flow to take effect. To achieve the best results, teams should adopt a "standardize-then-specialize" approach: establish global security and documentation rules at the user level, while using workspace-level `AGENTS.md` files to define the unique architectural patterns and tech stacks of individual projects.

Machine Learning Database Design TypeScript Gitlab Duo+4

gitlabJan 14, 2026Original article

Get started with GitLab Duo Agent Platform: The complete guide (opens in new tab)

The GitLab Duo Agent Platform represents a shift in AI-assisted development by moving from individual chat-based interactions to a collaborative multi-agent orchestration layer. By integrating specialized AI agents throughout the software development lifecycle, the platform transforms linear DevSecOps workflows into parallel processes that leverage full project context for tasks like security scanning and code refactoring. This architecture allows development teams to delegate routine technical burdens to autonomous agents, focusing human efforts on high-level innovation and complex problem-solving. ### Orchestrating the DevSecOps Lifecycle The platform functions as a central intelligence layer that connects AI agents to the broader GitLab ecosystem. * Agents access comprehensive project context, including source code management, CI/CD pipelines, issue tracking, and security scan results. * Specialized agents can be assigned to specific technical domains such as research, refactoring, and automated testing. * The system enables asynchronous collaboration, allowing multiple agents to work on different stages of a project simultaneously. ### Evolution from Duo Enterprise to Agentic AI The Duo Agent Platform is a superset of previous GitLab AI offerings, moving beyond simple 1:1 user-to-AI interactions. * GitLab Duo Pro focused on individual IDE productivity through code suggestions and basic chat. * GitLab Duo Enterprise expanded AI to the wider software lifecycle but remained primarily a 1:1 Q&A experience. * The Agent Platform introduces a many-to-many collaboration model where teams and multiple specialized agents interact autonomously to handle production-ready workflows. ### Advanced Integration and Customization To support enterprise-grade automation, the platform provides a roadmap for scaling AI from basic interactions to production environments. * Integration with the Model Context Protocol (MCP) allows for expanded data access and agent capabilities. * The platform supports a progression from initial agent interactions to full workflow customization and production-ready automation. * Developers can leverage the eight-part guide series to move from foundational concepts to advanced technical implementations. To maximize the benefits of agentic AI, organizations should transition from viewing AI as a simple Q&A tool to treating it as an orchestration layer. Teams are encouraged to explore the complete introductory series to begin delegating routine maintenance and security tasks to specialized agents, thereby accelerating overall delivery speed.

Machine Learning CI/CD Gitlab DevSecOps+3

gitlabJan 14, 2026Original article

Monitor, manage, and automate AI workflows (opens in new tab)

The GitLab Duo Agent Platform’s Automate capabilities provide a centralized framework for managing, executing, and monitoring AI-driven development workflows within the software development lifecycle. By integrating event-driven triggers and detailed session logging, the platform allows developers to transition from manual AI interactions to fully autonomous, production-ready processes. This orchestration layer ensures that AI agents are not only performant but also transparent and easy to audit across projects. ## Resource Management for Agents and Flows The Automate hub serves as the control center for organizing AI resources, distinguishing between agents (entities that perform tasks) and flows (structured sequences of actions). * Resources are categorized into "Enabled" (those available for project use) and "Managed" (those created and owned specifically by the project). * Custom agents and flows must be enabled at the top-level group before they can be activated for specific projects. * Users can expand their automation library by browsing and enabling pre-configured resources from the GitLab AI Catalog. ## Event-Driven Automation with Triggers Triggers allow AI agents to respond automatically to specific actions within the GitLab interface, eliminating the need for manual invocation. * Automation can be initiated through three primary event types: user mentions (e.g., `@agent-name`), issue/MR assignments, or reviewer assignments. * When a trigger is activated, the system identifies the associated flow, executes the agent, and posts the final results directly back to the relevant issue or merge request. * Common use cases include using the `/assign` quick action to trigger a CI/CD optimizer or a code explanation agent. ## Workflow Monitoring and Session Transparency The Sessions interface provides a detailed audit trail for every execution, offering visibility into the "black box" of AI decision-making. * The Activity tab tracks step-by-step reasoning, showing exactly which tools the agent used and the results of individual actions. * Execution statuses are monitored in real-time, with labels such as Running, Finished, Failed, or Input Required. * The Details tab provides deep technical context by linking directly to Runner job logs, including system messages and full tool invocation outputs. ## Practical Conclusion To maximize the utility of the GitLab Duo Agent Platform, teams should move beyond experimental chat prompts and begin configuring triggers for repetitive tasks like code review assignments or issue triaging. Utilizing the Sessions tool is recommended during the initial rollout phase to verify agent reasoning and ensure that custom flows are interacting correctly with project data before full-scale deployment.

Machine Learning Gitlab DevSecOps Event Driven Automation+4

googleJan 11, 2026Original article

NeuralGCM harnesses AI to better simulate long-range global precipitation (opens in new tab)

NeuralGCM represents a significant evolution in atmospheric modeling by combining traditional fluid dynamics with neural networks to solve the long-standing challenge of simulating global precipitation. By training the AI component directly on high-quality NASA satellite observations rather than biased reanalysis data, the model achieves unprecedented accuracy in predicting daily weather cycles and extreme rainfall events. This hybrid approach offers a faster, more precise tool for both medium-range weather forecasting and multi-decadal climate projections. ## The Limitations of Cloud Parameterization * Precipitation is driven by cloud processes occurring at scales as small as 100 meters, which is far below the kilometer-scale resolution of global weather models. * Traditional models rely on "parameterizations," or mathematical approximations, to estimate how these small-scale events affect the larger atmosphere. * Because these approximations are often simplified, traditional models struggle to accurately capture the complexity of water droplet formation and ice crystal growth, leading to errors in long-term forecasts. ## Training on Direct Satellite Observations * Unlike previous AI models trained on "reanalyses"—which are essentially simulations used to fill observational gaps—NeuralGCM is trained on NASA satellite-based precipitation data spanning 2001 to 2018. * The model utilizes a differentiable dynamical core, an architecture that allows the neural network to learn the effects of small-scale events directly from physical observations. * By bypassing the weaknesses inherent in reanalysis data, the model effectively creates a machine-learned parameterization that is more faithful to real-world cloud physics. ## Performance in Weather and Climate Benchmarks * At a resolution of 280 km, NeuralGCM outperforms leading operational models in medium-range forecasts (up to 15 days) and matches the precision of sophisticated multi-decadal climate models. * The model shows a marked improvement in capturing precipitation extremes, particularly for the top 0.1% of rainfall events. * Evaluation through WeatherBench 2 demonstrates that NeuralGCM accurately reproduces the diurnal (daily) weather cycle, a metric where traditional physics-based models frequently fall short. NeuralGCM provides a highly efficient and accessible framework for researchers and city planners who need to simulate long-range climate scenarios, such as 100-year storms or seasonal agricultural cycles. Its ability to maintain physical consistency while leveraging the speed of AI makes it a powerful candidate for the next generation of global atmospheric modeling.

Machine Learning AI Neural Networks Climate Modeling+5

figmaJan 7, 20262 min readCurated summary

Software Is Culture

Software has evolved from an invisible, mechanical tool into an intimate part of human experience. The next major shift is toward intelligent, adaptive systems that learn and change with users. The post argues that designers will shape this future not merely by automating tasks, but by deciding where technology directs human attention, curiosity, and creativity. ## Software as Human Experience - Software once operated behind screens as an efficient, obedient system. - Touch interfaces—pinching, swiping, and tapping—made software embodied and deeply connected to how people think, feel, and communicate. - As a result, user experience became inseparable from human experience. ## The Shift to Intelligent Software - Future systems will learn, adapt, and respond rather than follow fixed interactions. - Interfaces may change in real time based on users’ needs and behavior. - Static products will become fluid, adaptive, and increasingly “alive.” - Understanding this transition requires examining influential interactions and design decisions from the past two decades. ## AI as Tool and Collaborator - AI can amplify distinctly human qualities rather than simply replace human work. - Interaction is shifting from commands, layers, and parameters toward conversation, goals, gestures, and overall intent. - The central design question is not what to automate, but where people should focus their attention. ## The Designer’s Role - Designers are navigating how intelligent software will affect curiosity, meaning, and craft. - Their choices will influence how future generations experience and understand technology. - Finding a personal perspective as a designer is essential to shaping software’s cultural impact. The practical takeaway is to treat software design as cultural work: build adaptive systems thoughtfully, and focus on how they direct human attention and enable meaningful action.

Read original

Machine Learning AI User Experience Software Design+3

tossDec 22, 2025Original article

Toss's AI Technology Recognized (opens in new tab)

Toss ML Engineer Jin-woo Lee presents FedLPA, a novel Federated Learning algorithm accepted at NeurIPS 2025 that addresses the critical challenges of data sovereignty and non-uniform data distributions. By allowing AI models to learn from localized data without transferring sensitive information across borders, this research provides a technical foundation for expanding services like Toss Face Pay into international markets with strict privacy regulations. ### The Challenge of Data Sovereignty in Global AI * Traditional AI development requires centralizing data on a single server, which is often impossible due to international privacy laws and data sovereignty regulations. * Federated Learning offers a solution by sending the model to the user’s device (client) rather than moving the data, ensuring raw biometric information never leaves the local environment. * Standard Federated Learning fails in real-world scenarios where data is non-IID (Independent and Identically Distributed), meaning user patterns in different countries or regions vary significantly. ### Overcoming Limitations in Category Discovery * Existing models assume all users share similar data distributions and that all data classes are known beforehand, which leads to performance degradation when encountering new demographics. * FedLPA incorporates Generalized Category Discovery (GCD) to identify both known classes and entirely "novel classes" (e.g., new fraud patterns or ethnic features) that were not present in the initial training set. * This approach prevents the model from becoming obsolete as it encounters new environments, allowing it to adapt to local characteristics autonomously. ### The FedLPA Three-Step Learning Pipeline * **Confidence-guided Local Structure Discovery (CLSD):** The system builds a similarity graph by comparing feature vectors of local data. It refines these connections using "high-confidence" samples—data points the model is certain about—to strengthen the quality of the relational map. * **InfoMap Clustering:** Instead of requiring a human to pre-define the number of categories, the algorithm uses the InfoMap community detection method. This allows the client to automatically estimate the number of unique categories within its own local data through random walks on the similarity graph. * **Local Prior Alignment (LPA):** The model uses self-distillation to ensure consistent predictions across different views of the same data. Most importantly, an LPA regularizer forces the model’s prediction distribution to align with the "Empirical Prior" discovered in the clustering phase, preventing the model from becoming biased toward over-represented classes. ### Business Implications and Strategic Value * **Regulatory Compliance:** FedLPA removes technical barriers to entry for markets like the EU or Southeast Asia by maintaining high model performance while strictly adhering to local data residency requirements. * **Hyper-personalization:** Financial services such as Fraud Detection Systems (FDS) and Credit Scoring Systems (CSS) can be trained on local patterns, allowing for more accurate detection of region-specific scams or credit behaviors. * **Operational Efficiency:** By enabling models to self-detect and learn from new patterns without manual labeling or central intervention, the system significantly reduces the cost and time required for global maintenance. Implementing localized Federated Learning architectures like FedLPA is a recommended strategy for tech organizations seeking to scale AI services internationally while navigating the complex landscape of global privacy regulations and diverse data distributions.

Machine Learning AI Computer Vision Data Privacy+5