Yash Dewalkar

AI Engineer with hands-on experience developing Generative AI, RAG, LLM fine-tuning, Agentic AI, and Machine Learning solutions using Python, LangChain, LangGraph, Hugging Face, Qdrant, and PyTorch. Experienced in building production-oriented AI pipelines, semantic retrieval systems, LLM evaluation workflows, and REST API/Streamlit applications

Data Science Associate Intern - Darwinbox

(2023-09 - 2024-09)

• Evaluated and quality-validated 500+ LLM-generated responses from Llama2 and Mistral for relevance, factual consistency, and response quality.
• Implemented a Retrieval-Augmented Generation (RAG) pipeline for the Ask Darwin chatbot, improving retrieval relevance by approximately 12% through retrieval and prompt optimization.
• Conducted systematic LLM response evaluation and retrieval improvements, contributing to an estimated 15% reduction in incorrect or unsupported responses.
• Cleaned and pre-processed 20,000+ HR records for machine-learning model development, including missing-value treatment, feature preparation, and data-quality validation.
• Collaborated with cross-functional teams to design an Employee Attrition Prediction Model using machine learning techniques.

Bachelor of Engineering - Rajiv Gandhi College of Engineering & Research (2021) Nagpur, India

Post Graduate Program - Artificial Intelligence and Machine Learning - Great Learning (2022 - 2023)