About

I started my career building traditional software — ETL pipelines at Deloitte, fintech platforms in India, travel SaaS in Paris. Solid CRUD, clean architecture, AWS. The kind of systems that quietly run businesses.

Then the ground shifted. Today I work on Life Cycle Assessment software in climate tech, and my core mission is the most interesting engineering problem I've touched: re-architecting a traditional platform to be AI-native. Not “adding a chatbot” — rethinking the architecture: where retrieval replaces search, where agents replace forms, and where LLM pipelines need the same rigor — evals, observability, cost control — that we used to reserve for payments code.

I write about what I learn doing this: the architecture decisions, the failures, the benchmarks. If you're turning a traditional system into an AI-native one, follow along.

Elsewhere: GitHub · LinkedIn · dhanushk2105@gmail.com