Srinath Therampattil — Staff Engineer at Airbnb with 15 years building enterprise platforms. Writing on agentic AI, Salesforce platform engineering, and making AI reliable on systems of record.https://srinaththerampattil.com/https://srinaththerampattil.com/blog/two-agents-one-brief/https://srinaththerampattil.com/blog/two-agents-one-brief/I run Claude Code and Codex over the same repo, and each wants its own instructions file. Maintaining two by hand means they drift, and a stale project brief is worse than none. The fix is boring, and it's a symlink.Tue, 16 Jun 2026 09:30:00 GMThttps://srinaththerampattil.com/blog/how-claude-codes-creator-uses-it/https://srinaththerampattil.com/blog/how-claude-codes-creator-uses-it/Boris Cherny built Claude Code, and the way he uses it isn't about clever settings. It's about running it like a small team instead of a chat window. The handful of habits behind that, and the ones I've adopted.Tue, 16 Jun 2026 00:00:00 GMThttps://srinaththerampattil.com/blog/ai-tools-that-compound/https://srinaththerampattil.com/blog/ai-tools-that-compound/AI coding tools forget everything between sessions, so they repeat mistakes and relearn your conventions over and over. Here's the learnings loop that fixes it, and exactly how I wired it into Claude Code on this blog.Mon, 15 Jun 2026 00:00:00 GMThttps://srinaththerampattil.com/blog/putting-ai-in-front-of-a-platform/https://srinaththerampattil.com/blog/putting-ai-in-front-of-a-platform/What I've learned putting LLMs and agents on top of a Salesforce platform, where the data has rules you don't get to ignore and a wrong answer lands in the system the business runs on.Sun, 14 Jun 2026 00:00:00 GMThttps://srinaththerampattil.com/blog/reliable-ai-agents-on-enterprise-platforms/https://srinaththerampattil.com/blog/reliable-ai-agents-on-enterprise-platforms/An agent that can change records on a system of record is powerful and risky in equal measure. The guardrails I rely on — acting as the user, idempotent actions, a narrow toolset, human checkpoints, and real logging — to let one run safely.Sat, 13 Jun 2026 00:00:00 GMThttps://srinaththerampattil.com/blog/building-reliable-llm-features/https://srinaththerampattil.com/blog/building-reliable-llm-features/An LLM feature is easy to demo and hard to trust. These are the practical habits — validating output, measuring quality, versioning prompts, and planning for wrong answers — that I rely on to make one hold up with real users.Fri, 12 Jun 2026 00:00:00 GMThttps://srinaththerampattil.com/blog/nfrs-for-ai-systems/https://srinaththerampattil.com/blog/nfrs-for-ai-systems/Most teams spec what an AI feature should do and skip how well it has to do it. The non-functional requirements — accuracy, latency, cost, fallback, observability, governance — that decide whether it's actually production-ready.Thu, 11 Jun 2026 00:00:00 GMThttps://srinaththerampattil.com/blog/evaluating-llm-output-without-ground-truth/https://srinaththerampattil.com/blog/evaluating-llm-output-without-ground-truth/You almost never have labeled data when you ship an AI feature. A practical way to measure quality anyway — a small hand-built set, plain assertions, a checked model-as-judge, and the production signals you already have.Mon, 01 Jun 2026 00:00:00 GMT