Data Engineer at trigyan.io · previously AWS (DynamoDB)
> builds data platforms & AI systems that run themselves
I'm a data engineer at trigyan.io, building the data platform behind a healthcare product. Before this I was at Amazon Web Services on the DynamoDB team — backup, restore and import at multi-terabyte scale — and built Azure Databricks lakehouse pipelines in consulting.
What I like working on: pipelines that move serious data — Spark, Kafka, Databricks, medallion architectures — and LLM systems with real engineering discipline: deterministic where possible, model calls only where judgment helps, everything tested and observable. To practice what I preach, I built a fleet of autonomous agents that runs my mornings, audits my code, and even maintains my GitHub profile page by itself.
My learning isn't a private plan — it's committed, dated and public. A self-updating board tracks daily progress across four tracks, with code and day-by-day notes pushed as I go, alongside daily DSA practice. What's mine here is the progress; a Telegram study bot just keeps the log.
Twelve autonomous agents — each its own repo, schedule, Telegram bot and memory — built to run my mornings: mail triage, news, tech briefings, my own money movement, a code review of everything I pushed yesterday. A Cloudflare Worker dispatches GitHub Actions at exact minutes (after GitHub's own cron once ran 4 hours late); state is committed back after every run; a watchdog notices any agent that dies silently. The fleet scales up and down with my needs — and one agent audits every repo I own, with a live one-button dashboard.
Async FastAPI backend for news aggregation and querying — careful query design, async plumbing throughout.
Election-results dashboard with real data modeling and caching (TypeScript).
RAG document Q&A — parse, embed, ask. A compact end-to-end retrieval pipeline in Python.
Published research (IOP, 2020): YOLO object detection over Kafka-streamed video with text-to-speech, on Android.
“Real-time Object Detection and Face Recognition System to Assist the Visually Impaired” — Journal of Physics: Conference Series, Vol. 1706 (IOP Publishing), 2020 · ICAPSM 2020.
Have a role, a project, or just want to talk data platforms and agents? My inbox is open.