Backend and platform engineer focused on production reliability, multi-tenant systems, security, automation, and AI infrastructure.
I work on systems where correctness matters: database access, event-driven workflows, tenant isolation, observability, evaluation harnesses, and production debugging.
- production-reliability-lab — reproduce PostgreSQL connection exhaustion with synthetic data and measure six responses: app-side pool, pgbouncer, statement/idle-transaction timeouts, and the code change that removes the leak. Load generator,
pg_stat_activitycollector, charts with per-endpoint error attribution, incident runbook, CI. - telco-customer-churn-mlops — end-to-end MLOps pipeline: DVC + MLflow + Hydra, BentoML service, GitHub Actions → GHCR, drift monitoring. 38 tests (Pandera, Deepchecks, HTTP integration), leakage-free training.
- sensor-data-streaming-pipeline — Kafka + Schema Registry + MongoDB streaming pipeline with idempotent producers, at-least-once delivery, dead-letter queue, and retry with backoff. Dockerised, CI-tested.
- ocr-pipeline — benchmarking harness for OCR engines (EasyOCR, TrOCR, Tesseract): solver lifecycle, 11 preprocessing variants, accuracy and latency reports.
Production debugging and root-cause analysis · PostgreSQL security, pooling, and connection architecture · multi-tenant authorization and data isolation · event-driven automation with transactional outbox · performance work with measured before/after · TDD and adversarial review.
- Backend reliability and performance audits
- PostgreSQL connection, query, and access reviews
- Multi-tenant authorization reviews
- AI evaluation and agent-reliability harnesses
- Fixed-scope production debugging
📍 Bengaluru, India · Remote