Blog

Thoughts & Stories

Sharing my experiences, learnings, and adventures in tech and beyond.

Browse by Topic

DevOps & Cloud Infrastructure (143)AI & LLMs (130)Web Development (81)ERP Systems (74)Backend & APIs (64)Software Development (52)Security (30)Career Reflections (27)Cloud (16)Product & SaaS (12)AI / Machine Learning (5)University (1)Hackathons (1)

Cloud

October 202614 min read

Cloudflare Containers: 6x Faster Starts for Agent Sandboxes

Cloudflare rebuilt Containers for agent sandboxes: a 648 ms median start, runtime image and instance selection, filesystem snapshots, and a 31 December 2026 migration deadline.

A2A Protocol Tutorial: v1.0 Agent Cards, Tasks and Bindings

Backend

October 202612 min read

A2A Protocol Tutorial: v1.0 Agent Cards, Tasks and Bindings

A hands-on A2A protocol tutorial for v1.0: publish and sign an Agent Card, drive the task lifecycle, stream or push updates, choose a binding, and expose an ERP stock-check agent to a partner without the security gaps researchers found.

October 202612 min read

A2A vs MCP: Which AI Agent Protocol Do You Actually Need?

MCP connects an agent to tools; A2A connects agents to each other. Discovery, state, long-running tasks, auth and governance compared in one table.

October 202611 min read

AI Agent Cost Optimization: Budget Per Task, Not Per Call

Agents loop, spawn subagents and grow their context, so per-call pricing math underestimates them. Measure cost per completed task, then control it with max_turns, subagent caps, tool search, compaction and model routing.

October 202612 min read

AI Agent Evaluation: How to Test Tool Calls and Trajectories

An agent can return the right answer after calling the wrong tool, with the wrong arguments, in the wrong order. Here is how to grade the trajectory itself, from production traces to promptfoo assertions against a NestJS-hosted agent.

AI Agent Least Privilege: Credentials, Tokens and Secrets

Security

October 202611 min read

AI Agent Least Privilege: Credentials, Tokens and Secrets

Credential design for tool-using AI agents: keep secrets out of the context window, use domain_secrets in sandboxes, audience-bound per-request MCP tokens, one scoped identity per tool, and approval gates on writes.

Showing 1–6 of 636 posts

Explore More

Projects

See the projects I've built

Experience

My professional career

Certifications

Professional certifications