backend & platform engineer in bangalore. three years of distributed systems, database internals, and cloud infrastructure.
$ docker compose up -d # three nodes come up and elect a leader. $ curl -X PUT localhost:8080/kv/hello -d 'world' {"ok":true} $ docker compose kill node1 # leader gone. the other two re-elect. $ curl 'localhost:8081/kv/hello?consistency=linearizable' world
i own the backend deployment platform: github actions pipelines shipping docker images to AWS ECS across regions, plus a self-hosted mongoDB replica set, redis, rabbitmq, and a typesense search cluster.
i worked on a vector database in C++! read/write separation, a kafka-backed WAL, backup/restore, and HNSW index resets.
i migrated core microservices from flask to fastAPI for big latency wins, and built LLM + RAG workflows for a conversational AI agent.
i built a real-time CCTV video-analytics pipeline for an iDEX defense project and ran the microservices on amazon ECS.
i like systems problems: database internals, distributed consensus, and making slow APIs fast.
WALs, replication, elections, and the p99 that nobody can explain.
i enjoy owning things end-to-end, from the fastAPI service all the way down to the CI/CD pipeline and the AWS footprint it runs on. if it pages at 3am, i want to know how it works.
i'm an indian citizen based in bangalore, with work authorization in india. you can reach me at burhankapdawala[that little symbol that usually goes here]icloud.com!
a fault-tolerant distributed key-value store in C++ and go: raft consensus, gRPC, and WAL durability. consensus is fun when it's your own nodes disagreeing!