I've built backend and production infrastructure systems at WebEngage, Netomi, BluSmart, and MX Player, spanning access services, observability tooling, Kubernetes delivery, and automated recovery workflows.
At Netomi, I built FastAPI and Celery model-evaluation services, fine-tuned a Llama-3 function-calling agent, and developed multi-tenant retrieval backends using Milvus and Qdrant. My work reduced evaluation runtime, retrieval latency, and GPU inference response time while supporting enterprise AI workloads.
I work across Python, Go, AWS, GCP, Kubernetes, CI/CD, monitoring, testing, and applied AI, with experience turning evaluation, security, and operational requirements into production systems.
