Model design, training, evaluation and deployment for classification, prediction and generative workflows.
LoRA/PEFT fine‑tuning, SFT, RAG, embeddings and prompt engineering for robust LLM applications.
Invoice and document pipelines (OpenCV, Azure Document Intelligence) with proven extraction accuracy.
FastAPI, Docker, CI/CD, monitoring, vector DBs and cloud deployments (Azure/AWS).
Rapid prototyping, stakeholder demos, and productization from prototype to MVP.