Software Engineer
Current• Designed and developed context indexing and RAG (Retrieve, Augment, and Generate) pipelines, enabling business users to create Large Language Model (LLM) solutions by dragging UI components of the SnapLogic platform.• Developed the LLMs-related components using Java, allowing users to interact with embedding models and chat completions from OpenAI, Azure OpenAI, AWS Bedrock ( Titan and Claude) and configure the settings.• Implemented a Chatbot UI using Next.js, React, TypeScript and built the CI/CD using Docker, and GitHub Actions, enabling both internal and external users to engage in dialogues with document-based data.•Researched emerging LLM frameworks and developed interactive demos using Langchain, Python and Pinecone, including a novel chunking strategy for original text retrieval and a prompt evaluator to measure response accuracy of LLMs via an accuracy score.