Building context-aware applications using Large Language Models, prompt engineering, and custom vector search. Experienced in training, fine-tuning, and optimizing vision/speech pipelines for production use.

• InsightFace embedding similarity scoring
• YOLOv8/v9 object and person detection models
• Real-time speech translation and VITS (TTS) synthesis
• RAG pipeline creation and semantic indexing

Architecting high-performance asynchronous backends and server structures. Specialized in WebSocket connections for real-time bi-directional streaming, RESTful API design, and communications triggers.

• FastAPI async framework development
• Full-duplex WebSocket communication streams
• Twilio API integration for voice-calling agents
• Asynchronous messaging pipelines and logging systems

Designing databases for AI storage, semantic matching, and analytical queries. Experienced in configuring relational databases and vector indexing engines for rapid cosine-similarity lookups.

• Qdrant vector collection setup and queries
• PostgreSQL database schemas and connection pooling
• NumPy & Pandas for mathematical/data operations
• SQLAlchemy ORM integrations

Deploying and managing microservices in production environments. Optimizing deep learning models for CPU/GPU runtimes and setting up automated CI/CD deployment pipelines.

• Docker containerization for portable runs
• ONNX Runtime conversion for optimized AI inference
• AWS service deployment and management
• GitHub Actions workflows and Vercel hosting