● Collaborated with 2 clients to develop and deploy LLM-based chatbots using LangChain and Hugging Face Transformers, improving customer support response time by 40% . ● Built a fully autonomous agentic system using LangGraph, LangChain, LangSmith, CrewAI, Pinecone, and Vector DBs for multi-step reasoning and decision-making tasks. ● Delivered a credit card fraud detection API using FastAPI and AWS Lambda; achieved 93% accuracy through SMOTE balancing and feature optimization. ● Evaluated and optimized ML models with MLflow, boosting F1-score from 0.78 to 0.91. Provided expertise in AI inference and MLOps deployment using Docker, GitHub Actions, AWS S3, and Glue. Documented processes and presented results to stakeholders.

● Designed and implemented a Retrieval-Augmented Generation (RAG) chatbot using LangChain and Phi Data, leveraging NLP and ML techniques to enhance conversational accuracy and contextual relevance. ● Integrated the chatbot with OpenAI GPT models and deployed it via Streamlit on a web platform, enabling real-time interaction and scalable user support. ● Managed dynamic knowledge base integration using LangSmith, ensuring up-to-date and consistent chatbot responses across queries. ● Improved response quality and user engagement by over 25% through prompt refinement and iterative testing.Applied AI Ethics principles in LLM usage, prompt tuning, and bias mitigation.