Satyam Shree

I've built and optimized distributed data platforms at CAST Software, where I design versioned ETL workflows for historical analysis, incremental updates, and reliable state reconstruction.

I reduced ETL imports from 20+ hours to 7 hours and cut a customer-linking pipeline from 19 hours to 2.5 hours across datasets with 1M+ nodes and 2.2M+ caller-callee pairs. I also reduced excluded-link counting from nearly 14 hours to about 15 minutes and ETL startup time from 21 minutes to under one minute.

I work primarily in Go, SQL, PostgreSQL, Redis, Neo4j, gRPC, and event-driven architectures. My work includes resolving production deadlocks and race conditions in multi-tenant job queues, improving failure recovery, and building Python MCP servers for LLM-driven ETL execution.

I've also built a distributed Go pub/sub broker and a Go microservice URL shortener, with Redis, PostgreSQL, Docker, and Kubernetes support. Earlier, I developed production backend services with Django, RabbitMQ, Celery, Python, JavaScript, and AWS-integrated workflows.