• Build LLM-powered systems (RAG, agents, copilots, automation tools) • Design data pipelines for structured and unstructured data (PDFs, APIs, databases) • Implement retrieval systems (vector DBs, embeddings, semantic search) • Integrate and evaluate models (OpenAI, Claude, Gemini, open-source) • Ship APIs and microservices (FastAPI / Flask / Node) • Implement prompt engineering, evaluation, and guardrails • Deploy on cloud (AWS / GCP / Azure / IBM) • Monitor accuracy, latency, cost, and failure modes • Work directly with stakeholders to turn fuzzy needs into shipped systems