Senior Data Engineer based in Noida, India. Around twelve years building data platforms and pipelines.
Mostly Python, SQL and Spark. Day to day that means Databricks, Snowflake, Airflow and dbt, across Azure, AWS and GCP. Lately a lot of agentic AI and LLM plumbing on top of the same.
Currently reading the Azure file IO paths in PyIceberg and Airflow's Azure provider, with a view to contributing there.
killcount — an agent that writes tests and then proves they actually catch bugs. 666 injected faults, four different approaches, and a metric that rewards writing the bug in rather than writing around it.
gaze-mouse-kvm — a software KVM that moves your cursor between two machines using webcam gaze and head tracking.
ai_candidate_ranker — heuristic candidate ranking that scores 100k profiles in about 8 seconds on CPU. No embeddings, no ML dependencies.
BabelsGlyph — a side scrolling endless runner built with pygame-ce. Plays in the browser through WebAssembly.
Open to talking about data platform work, lakehouse table formats and pipeline reliability.
