Debojyoti Bhuinya

I’m an AI/ML Engineer focused on building reliable, scalable backend systems that power intelligent, AI-driven applications. In my current role, I designed and deployed scalable backend services for an AI-driven tutoring platform, handling real-time request processing and asynchronous workflows with Python APIs. I built production-grade CI/CD pipelines (GitHub Actions + AWS CLI) to auto-scale ML servers across dev/prod/demo environments, cutting deployment cycle time by ~60%, and I engineered ETL and data processing pipelines to ingest, transform, and serve structured and unstructured data for RAG-based systems.

I also take observability and performance seriously: I implemented a full observability stack (Prometheus, Grafana, Loki, OpenTelemetry) for monitoring system health, request tracing, and alerting, then optimized backend performance to reduce response latency for real-time AI interactions. Earlier, I delivered RAG-based conversational AI with ChromaDB and AWS Bedrock (including WebSocket-based real-time communication), and I developed deep learning models (YOLOv8, speaker diarization) for production via Hugging Face and Gradio. From applied research and conference publications to building multimodal detection platforms and production APIs, my ethos is the same—ship dependable AI systems with strong engineering foundations, from data pipelines to deployment and monitoring.