Generative AI Engineer
SDK Consulting Inc · Feb 2025 – Present
- Engineered advanced RAG chatbot with Milvus vector database, achieving 10X token throughput using SGLang and asynchronous processing.
- Implemented agentic AI for intelligent metadata filtering and recursive retrieval to maximize context relevance.
- Reduced hallucinations through prompt engineering constrained to retrieved context, and built document summarization with pre-indexed Q&A capabilities.
- Deployed scalable architecture using Ray Serve for multi-client inference, LlamaIndex for rapid development and model flexibility, with containerized CUDA libraries for seamless deployment via Streamlit UI.













