Alex Merced's Data, Dev and AI Blog

Topic

AI agents

75 posts tagged “AI agents”.

  1. 30 min read

    MCP in Regulated Environments: What the Toronto MCP Dev Summit Agenda Says About Production Agents

    Registries, gateways, delegated identity, and audit: MCP Dev Summit Toronto 2026 patterns for production agents over lakehouse data.

  2. 30 min read

    Your Data Agents Are CPU-Bound, and the Hardware Market Just Noticed

    Analytics agents spend most of their time on CPU-bound tool calls, not the model. Size three planes separately and measure your ratio from traces.

  3. 32 min read

    How to Build a Golden-Question Benchmark for Your Data Agents

    Build a golden-question benchmark for data agents: frozen Iceberg tags, execution-based scoring with Arrow, and CI gates.

  4. 30 min read

    When a Data Agent Gives a Wrong Number: Incident Response for Agentic Analytics

    A plan for when a data agent gives a wrong number: severity levels, containment, Iceberg time travel reproduction, and blast radius queries.

  5. 30 min read

    Multi-Tenant Data Agents: Serving Many Customers From One Lakehouse

    Serve many customers from one lakehouse without leaking between them: isolation models, Polaris grants, and tenant-scoped agent state.

  6. 30 min read

    Release Management for Data Agents: Shipping Prompts, Tools, Models, and Metrics Without Breaking Answers

    Ship prompts, tools, models, and metric changes without breaking answers: agent manifests, shadow mode, canaries, and rollback.

  7. 31 min read

    Semantic Caching for Data Agents Without Serving Stale or Leaked Numbers

    Cache data agent results safely with canonical query keys, Iceberg snapshot freshness checks, and permission-aware cache keys.

  8. 31 min read

    Small Language Models for Data Agents: Where They Win and Where They Fail

    Where small language models win inside a data agent and where they fail, plus a routing architecture that escalates only when needed.

  9. 30 min read

    Turning an Analytics Question Into a Verified Agentic Graph

    A complete AGS 1.0 graph for governed metric questions, with verification gates, deterministic check scripts, and reconciliation against a semantic layer.

  10. 30 min read

    Twelve Queries That Pass Every Schema Check and Still Return the Wrong Number

    Twelve query patterns that pass schema validation and still return wrong numbers, with detectors, required metadata, and tested code for each.

  11. 31 min read

    How Apache Ossie Is Deciding What Agents and BI Tools Can Ask a Semantic Layer

    How Apache Ossie's layered query design gives AI agents both a constrained dimensional interface and a grain-safe SQL interface for semantic layers.

  12. 30 min read

    Securing the Agentic Lakehouse Gateway: Preventing Prompt Injection and Data Exfiltration

    Agentic lakehouse gateways face prompt injection and exfiltration through query results. A threat model and defenses for the layer in front of data.

  13. 31 min read

    Metric Contracts in Code: Testing, Versioning, and Serving Business Logic to Multi-Agent Systems

    Metric contracts in code let teams test, version, and serve business logic to multi-agent systems without each agent inventing its own SQL.

  14. 30 min read

    Query Routing at Machine Scale: Dynamic Workload Distribution Across Lakehouse Engines

    Route each lakehouse query by shape, not by sender. Signals, rules, and how to keep dashboards, batch jobs, and agents from sharing one engine.

  15. 35 min read

    The Agent Is Now a Named Coworker, and It Needs a File Format

    Named, persistent agents need a file format. Open Agent Profile, Buzz, Grok Bot, and Hermes Bot Mode show why a portable agent identity matters.

  16. 32 min read

    Your Agent Should Answer the Phone: A Field Guide to AI Gateways on Slack, Discord, Telegram, Signal, and Teams

    A field guide to AI gateways on Slack, Discord, Telegram, Signal, and Teams: architecture, auth, cost, and the failure modes that matter.

  17. 32 min read

    Graphs in AI Engineering Have Solved Three Problems. The Fourth Is the Plan.

    Knowledge graphs, GraphRAG, and LangGraph solved three problems. The fourth is the work itself: a reviewable graph of bounded agentic loops.

  18. 30 min read

    Goal-Directed Data Quality Agents: Anomaly Quarantine on Apache Iceberg

    Goal-directed data quality agents that watch Apache Iceberg tables, detect anomalies, and quarantine suspect data safely with snapshot isolation and branches.

  19. 30 min read

    Metric Contracts in 2026: Standardizing Business Logic Across Multi-Agent Frameworks

    Metric contracts in 2026: versioned, testable definitions of business logic that let multi-agent frameworks compute revenue identically, with OSI interchange.

  20. 31 min read

    Building Stateless AI Tool Gateways with FastMCP, the 2026 MCP Spec, and Kubernetes

    How to build stateless AI tool gateways with FastMCP, the 2026 MCP specification, and Kubernetes, and why statelessness finally makes MCP scale.

  21. 23 min read

    The Plan and the Worker: Two Open Specifications for Agent Harnesses

    Two open specifications, the Agentic Graph Specification and the Open Agent Profile, turn agent plans and agent identity into portable, reviewable files.

  22. 21 min read

    Budgeting for Agentic Analytics When Every Question Costs Something Different

    Budgeting for agentic analytics when every question costs something different: token economics, query economics, instrumentation, and the cost controls.

  23. 20 min read

    The Five Layers of an Agentic Lakehouse and Where the MCP Server Sits

    The five layers of an agentic lakehouse and where the MCP server sits: storage, catalog, semantic layer, MCP gateway, and agent surface, plus identity.

  24. 21 min read

    Wiring Analytical Queries to Transactional APIs in Closed-Loop Decision Agents

    Wiring analytical queries to transactional APIs in closed-loop decision agents: conditional writes, sagas with compensations, decision records, and blast.

  25. 20 min read

    Why Agentic AI Needs a Governed Semantic Layer Behind the Model Context Protocol

    Why agentic AI needs a governed semantic layer behind the Model Context Protocol: metric consistency, access control, Apache Ossie for portable.

  26. 20 min read

    Moving From Supply Chain Dashboards to Decision Loops With the Model Context Protocol

    Moving from supply chain dashboards to decision loops with MCP: sense, decide, act, and verify, with typed action tools, idempotency keys, and graduated.

  27. 20 min read

    Metric Contracts as the Interface AI Agents Actually Need

    Metric contracts as the interface AI agents need: calculation, inclusion rules, grain, temporal semantics, ownership, semantic versioning, and testing.

  28. 20 min read

    Defending the Lakehouse Gateway Against Prompt Injection and Data Exfiltration

    Defending the lakehouse gateway against prompt injection and data exfiltration: per-user identity, no-SQL tool surfaces, volume bounds, and detection.

  29. 31 min read

    Guardrails for Analytics Agents That Do More Than Answer Questions

    The risk isn't agents going rogue, it's agents acting correctly on bad input at machine speed.

  30. 31 min read

    Building Agent Telemetry Tables in Iceberg That Survive an Audit

    A practical guide to building agent decision traces in Apache Iceberg that support audit reconstruction, governance review, and cost attribution.

  31. 30 min read

    What Agentic Analytics Actually Costs, and How to Keep It Bounded

    Agent analytics generates two cost streams that scale on different variables. Here's the arithmetic, the levers that actually move the number, and how.

  32. 30 min read

    When the Query Optimizer Starts Managing Its Own Materializations

    Autonomous materialized view management replaces quarterly review meetings with workload-driven scoring, and it's essential when AI agents generate.

  33. 30 min read

    Why AI Agents Fail on Raw Data, and What to Give Them Instead

    Agents fail on raw lake data because business rules live in people's heads. Data products with semantic contracts fix this at the source.

  34. 30 min read

    The Five Layers Between Your Lakehouse and a Trustworthy Agent

    Agent reliability is a property of the stack the model sits on. Five layers with distinct owners and failure modes turn the agent is unreliable.

  35. 31 min read

    Surviving Commit Conflicts When Dozens of Writers Hit the Same Iceberg Table

    Commit conflicts multiply with writer count, and AI agents introduce unpredictable write patterns.

  36. 31 min read

    Wiring an AI Agent to Apache Polaris with the Model Context Protocol

    The catalog is the right attachment point for AI agents working against a lakehouse. Here's how to wire the official Polaris MCP Server and add the read.

  37. 30 min read

    Governing What Agents Cost You

    Agents break the four assumptions analytics platforms were built on. A practical guide to identity, budgets, semantic layers, caching, and instrumentation.

  38. 20 min read

    Active Analytics Loops for AI Action Agents

    An in-depth exploration of active analytics loops for ai action agents

  39. 18 min read

    The Chase-Catch Gap in Enterprise AI Agents

    An in-depth exploration of the chase-catch gap in enterprise ai agents

  40. 18 min read

    Data Lakehouse Open Standards for AI Agents

    An in-depth exploration of data lakehouse open standards for ai agents

  41. 18 min read

    Five-Layer Agentic Lakehouse Architecture

    An in-depth exploration of five-layer agentic lakehouse architecture

  42. 18 min read

    GSA MCP Servers for Open Data AI Agents

    An in-depth exploration of gsa mcp servers for open data ai agents

  43. 18 min read

    Iceberg Concurrency for AI Agent Writes

    An in-depth exploration of iceberg concurrency for ai agent writes

  44. 17 min read

    MCP Stateless Gateways for Data Agents

    An in-depth exploration of mcp stateless gateways for data agents

  45. 17 min read

    Semantic View Autopilot for AI Governance

    An in-depth exploration of semantic view autopilot for ai governance

  46. 15 min read

    Building Closed-Loop Decision Agents: Moving from Passive BI Dashboards to Active Goal-Directed Workflows

    Dashboards are excellent at showing people what happened. They are less good at deciding what should happen next. That gap is where closed-loop de...

  47. 26 min read

    Designing Your Own AI Harness: A Deep Dive Into the Architecture of Agent Loops, Tools, Context, and Control

    A deep dive into custom AI harness architecture: model layers, tool design, context management, permissions, control budgets, persistence, orchestration.

  48. 26 min read

    Deterministic Data Engineering With AI Harnesses: Using Claude Code, Codex, Antigravity, and OpenCode for Data Work You Can Actually Trust

    How to use AI agent harnesses for data engineering without losing determinism, reproducibility, and trust in your data pipelines and analytics.

  49. 15 min read

    Enforcing Fine-Grained Security at Machine Speed: Dynamic Access Control for High-Frequency AI Agents

    AI agents change the security model for analytics. A human user may run a handful of queries, pause, interpret the answer, and ask a follow-up. An...

  50. 15 min read

    The Metric Contract Mandate: Standardizing Semantic Layers Before AI Agent Access

    AI agents are very good at moving quickly. That is the opportunity and the risk. If an agent can inspect metadata, generate queries, compare result...

  51. 30 min read

    Personal Context vs. Shared Context: A Deep Dive Into How Humans and Organizations Should Feed Their AI Agents

    The most important discovery of the agent era fits in one sentence: most AI failures are context failures, not model failures. When your assistant...

  52. 14 min read

    Autonomous Materialization for Agentic Analytics

    Autonomous materialization is useful when it is tied to workload evidence, governance checks, and lifecycle management.

  53. 15 min read

    Composable Semantic Layers for Analytical Agents

    AI agents need more than metric names. They need composable business logic that survives multi-step analysis.

  54. 14 min read

    The Context Layer for AI Agents

    A semantic layer is necessary, but agents also need lineage, quality, freshness, compliance, and ownership context.

  55. 14 min read

    Fine-Grained Security for AI Agents

    Machine-speed analytics requires machine-enforced policy, identity, masking, filtering, and audit controls.

  56. 34 min read

    The Complete Guide to Agentic Coding Tools in 2026

    Compare Codex, Claude Code, OpenCode, Gemini CLI, and GitHub Copilot CLI by workflow, model choice, governance, and cost controls.

  57. 14 min read

    Agentic Lakehouse Concurrency and Isolation

    How Iceberg optimistic concurrency control, partition-level isolation, and idempotency keys enable safe concurrent writes from multiple AI agents.

  58. 14 min read

    Anatomy of an Agentic Lakehouse

    The four-layer architecture of the agentic lakehouse: object storage, Apache Iceberg table format, Apache Polaris catalog, and the semantic/agent layer.

  59. 14 min read

    Real-Time Agentic Analytics with ClickHouse

    ClickHouse has become the leading real-time analytics engine for AI agent workloads, with event-loop architectures that let agents query and act.

  60. 14 min read

    Goal-Directed Analytics Agents on Apache Iceberg

    How goal-directed analytics agents decompose business questions into sub-tasks, execute action loops over Apache Iceberg tables, and use the lakehouse.

  61. 14 min read

    Microsoft Fabric Build 2026 Agentic Analytics Stack

    Microsoft Build 2026 revealed an agentic analytics stack built on Fabric IQ, OneLake Iceberg support, and semantic models.

  62. 14 min read

    Securing Agent Identities in the Lakehouse

    How OAuth 2. 0 token exchange, OAuth 2. 1 device authorization grant, credential vending, and fine-grained access control secure AI agent identities.

  63. 16 min read

    Data Platform Native AI Agent Tooling in 2026

    A comprehensive comparison of AI agent tooling across Dremio, Snowflake, Databricks, Microsoft Fabric, AWS, Google Cloud, ClickHouse, VeloDB, SpiceAI.

  64. 7 min read

    Use Hermes Agent for Free With DeepSeek V4 and Slack

    Install Hermes Agent, choose a current DeepSeek provider, and connect Slack through Socket Mode with the tokens, scopes, events, and allowlist Hermes requires.

  65. 5 min read

    A Journey from AI to LLMs and MCP - 10 - Sampling and Prompts in MCP – Making Agent Workflows Smarter and Safer

    Sampling and Prompts in MCP – Making Agent Workflows Smarter and Safer

  66. 5 min read

    A Journey from AI to LLMs and MCP - 9 - Tools in MCP – Giving LLMs the Power to Act

    Tools in MCP – Giving LLMs the Power to Act

  67. 5 min read

    A Journey from AI to LLMs and MCP - 8 - Resources in MCP – Serving Relevant Data Securely to LLMs

    Resources in MCP – Serving Relevant Data Securely to LLMs

  68. 5 min read

    A Journey from AI to LLMs and MCP - 7 - Under the Hood – The Architecture of MCP and Its Core Components

    Under the Hood – The Architecture of MCP and Its Core Components

  69. 5 min read

    Journey from AI to LLMs and MCP - 6 - Enter the Model Context Protocol (MCP) – The Interoperability Layer for AI Agents

    Enter the Model Context Protocol (MCP) – The Interoperability Layer for AI Agents

  70. 8 min read

    A Journey from AI to LLMs and MCP - 5 - AI Agent Frameworks – Benefits and Limitations

    AI Agent Frameworks – Benefits and Limitations

  71. 5 min read

    A Journey from AI to LLMs and MCP - 4 - What Are AI Agents – And Why They're the Future of LLM Applications

    What Are AI Agents – And Why They're the Future of LLM Applications

  72. 5 min read

    A Journey from AI to LLMs and MCP - 3 - Boosting LLM Performance – Fine-Tuning, Prompt Engineering, and RAG

    Boosting LLM Performance – Fine-Tuning, Prompt Engineering, and RAG

  73. 5 min read

    A Journey from AI to LLMs and MCP - 2 - How LLMs Work – Embeddings, Vectors, and Context Windows

    How LLMs Work – Embeddings, Vectors, and Context Windows

  74. 4 min read

    A Journey from AI to LLMs and MCP - 1 - What Is AI and How It Evolved Into LLMs

    What Is AI and How It Evolved Into LLMs

  75. 15 min read

    Building a Basic MCP Server with Python

    The Basics of Building a Basic MCP Server

Browse all posts

Work with Alex

Book Alex to speakGet the newsletterFree Apache Iceberg book

Menu

Search

Type at least two characters.