Oct 2021 – Present · 4 yrs 7 mos · Kolkata · On-site

• Deployed and configured open-source Large Language Models (LLMs) (e.g., Llama 3, Mistral) within a virtualized or cloud environment (Docker/AWS/GCP).
• Set up and initialized GenAI development environments using Python, focusing on libraries like Transformers (Hugging Face) and LangChain.
• Processed and curated specific domain data to create high-quality datasets for LLM fine-tuning or Retrieval-Augmented Generation (RAG) implementation.
• Implemented vector databases (e.g., Pinecone, ChromaDB) to store and retrieve document embeddings for contextual grounding in GenAI applications.
• Designed and iterated on advanced prompt engineering strategies to optimize the output and behavior of LLMs for specific tasks (e.g., summarization, code generation).
• Conducted low-rank adaptation (LoRA) fine-tuning on pre-trained LLMs to adapt them for a specialized industry domain (e.g., healthcare, finance).
• Built and tested an end-to-end application using a framework like Streamlit or Gradio to provide a user interface for the LLM.
• Implemented a Retrieval-Augmented Generation (RAG) pipeline using LangChain or LlamaIndex to enable the LLM to generate context-aware responses from external knowledge.
• Managed the project codebase and tracked changes using Git and GitHub, adhering to standard software development practices.
• Evaluated model performance and cost implications of various LLM APIs (e.g., OpenAI, WatsonX) and deployment options.