At EROS GenAI, I build the Eros Universe AI Tools Platform: seven production microservices delivering image, video, voice, and avatar generation through FastAPI, GPU workers, Kafka, Redis, PostgreSQL, and Kubernetes.
I've delivered end-to-end generative AI tools and redesigned platform-wide credit deduction and refund flows so charges occur only after moderation and successful queueing. I also standardized 80+ machine-readable error codes and integrated content moderation across media-generation workflows.
I solve distributed-systems reliability problems, from event-driven Kafka migrations and Redis-backed SSE fan-out to connection-pool exhaustion, event-loop blocking, and reliable GPU job processing. My scalability redesign removed a 50-stream-per-pod ceiling, while load and regression benchmarks gave architecture reviews measurable capacity and latency evidence.
Previously at Intellecto Global Services, I built Django and generative AI applications for document intelligence, healthcare, ERP, and land-registry workflows. My work reduced document-analysis time by 70%, automated 1,000+ daily ID verifications, and improved a healthcare API from 800ms to 180ms.
