Prathamesh Shete

Designed and deployed an LLM-driven interviewer system using LangChain, Redis, and Python for adaptive conversational evaluation. Built Transformer-based pipelines for real-time question generation and semantic analysis. Developed modular FastAPI microservices enabling scalable LLM inference and seamless REST integration. Utilized AWS GPU instances to optimize model inference speed and reduce latency. Collaborated closely with backend and frontend teams to integrate AI workflows into production systems.

Built RAG pipelines using Hugging Face Transformer models to improve factual grounding and retrieval quality. Engineered data preprocessing workflows for large-scale text and unstructured datasets. Performed evaluation and fine-tuning for generative tasks using BLEU, ROUGE, and accuracy metrics. Contributed to containerization and reproducible experiment setups using Docker.

Automated data extraction using BeautifulSoup and Selenium, generating structured knowledge representations. Supported the development of a Generative chatbot using FAISS and RAG for document-based question answering. Ensured high-quality data labeling and source validation to maintain dataset integrity.