At Socialsense, I built blue-green CI/CD pipelines for Azure App Services and asynchronous Python workers using RabbitMQ. I also led the technical design of the data intelligence platform and managed its development from ingestion to API deployment.
At SpeakX, I migrated 10+ REST API endpoints from Node.js to Golang, reducing average response latency by 40%. I also implemented gRPC-based audio streaming that reached ~800 ms to first audio byte, and an ETL pipeline that processed 180+ million records.
On my SignSpeak project, I engineered a model for extracting glosses from raw videos and improved sign language recognition accuracy by 79%. I also worked on contrastive learning and brain MRI segmentation projects using PyTorch.
