At Epoch Limited, I contributed to LUMINAIRE AI Conversation Analytics as call processing grew from about 3,000 to 25,000 calls a day. I developed Python components for audio ingestion, WhisperX transcription, LLM insights, and MongoDB persistence.
I modernized distributed processing with RabbitMQ workers, durable job states, retries, and domain prioritization, and contributed to multi-server deployments and vLLM integration. I also investigated ingestion and scheduling failures and contributed to Flask APIs, migrations, reconciliation, and production debugging.
As a freelance backend and automation developer, I design and maintain FastAPI services and Selenium automations for logistics, finance, and document platforms. I also build computer-vision projects, including YOLOv8 video analytics for road inspection, traffic, pedestrian movement, construction safety, and football.
