Posts

All the articles I've posted.

• Agentic Software Engineering — How I Build Production Systems in 2026

  Posted on:December 22, 2025 |  at 10:00 AM

  A practical account of how AI agents have changed the software engineering workflow in 2026 — from spec writing to code generation, automated testing, deployment, and production monitoring. What changed, what didn't, and what the transition felt like.

• Multi-Agent Workflows with Claude API — Architecture Patterns That Work

  Posted on:December 8, 2025 |  at 10:00 AM

  Production patterns for orchestrating multiple AI agents: parallel execution, hierarchical orchestration, agent-to-agent communication, and error recovery. Includes complete TypeScript implementations for real-world multi-agent workflows.

• TensorFlow.js vs scikit-learn in the Browser — Two Paradigms of Client-Side ML

  Posted on:November 17, 2025 |  at 10:00 AM

  TensorFlow.js brings neural networks to the browser natively. scikit-learn runs via Pyodide WebAssembly. Same Iris dataset, same task — totally different philosophies. Train both client-side with no backend and compare accuracy, interpretability, and the cold-start gap.

• Python vs JavaScript DataFrames in the Browser — Live Benchmarks with No Backend

  Posted on:October 27, 2025 |  at 10:00 AM

  Both Python (pandas via Pyodide WebAssembly) and JavaScript (arquero) can process DataFrames entirely in the browser. This post runs the same groupby, filter, and pivot benchmarks in both — live, client-side, no server needed — and measures the real tradeoffs.

• Building MCP Servers — The New API Layer for AI Agents

  Posted on:October 13, 2025 |  at 10:00 AM

  Model Context Protocol (MCP) is Anthropic's open standard for connecting AI models to external tools and data. This post builds a production-ready MCP server that exposes database queries and API calls as tools for Claude agents.

• Kubernetes for Backend Engineers — Pods, Deployments, and Services Without the Jargon

  Posted on:September 10, 2025 |  at 09:00 AM

  Kubernetes looks intimidating until you understand the one mental model that explains everything: declare desired state, and the control plane reconciles reality toward it continuously. A practical guide for backend engineers who deploy APIs and don't want to think about servers.