Research statement

I work at the intersection of Generative AI and systems, focusing on:

• LLM adaptation and alignment: LoRA/PEFT, RLHF/DPO/GRPO, quantization, safety and evaluation
• Agent systems: tool-use, planning and retrieval orchestration (LangChain, LangGraph, CrewAI, MCP)
• Retrieval and data engineering: advanced RAG with hybrid/vector search (ChromaDB/Milvus/Pinecone), knowledge graphs, high-quality pipelines
• Trustworthy ML: observability, robustness, privacy-aware deployment, and human-in-the-loop feedback

My goal is to build reliable AI systems that translate research into measurable outcomes in production while maintaining responsible practices.

I’m open to MS/PhD research collaboration in applied LLMs/agents, trustworthy ML, and data-centric AI.

Selected ongoing work

1) Transfer learning for bone marrow classification (medical imaging) 2) Thermal comfort modeling with ML and explainability across geographies 3) Federated learning in smart buildings for energy efficiency and comfort