Mayur Sai Kollipara

- A RAG application allowing users to chat with PDF documents, built with Streamlit and LangChain.

- Features a multi-model approach to compare responses from LLMs like Llama 3 and Amazon Nova Lite.

- Leverages Amazon Titan for high-quality text embeddings and FAISS for efficient similarity searching.

- Processes complex PDFs by partitioning them into distinct text, tables, and image blocks.

- Generates searchable AI summaries for all visual content, making charts and images fully queryable.

- Provides factually-grounded answers by reasoning across retrieved text, tables, and the original images simultaneously.

- A hybrid RAG system combining Neo4j Knowledge Graphs with Vector Search for deep, connected data retrieval.

- Uses Groq for high-speed inference and Ollama for local embeddings, turning Wikipedia topics into a searchable intelligence hub.

- Synthesizes answers by combining direct facts from the graph with semantic context from vector search.

- Multi-Agent Coding System: A sophisticated coding agent built with LangGraph that utilizes a three-tier architecture—Planner, Architect, and Coder agents—to automate software development.

- Automated Project Generation: Transforms high-level user task descriptions into fully functional codebases, delivering the final output as a downloadable ZIP file directly to the user's PC.

- Scalable Cloud Backend: Features a robust backend built with FastAPI and deployed on AWS EC2, ensuring efficient processing of complex coding tasks.

- Real-time Progress Tracking: Provides a Streamlit-based frontend that displays a live execution trace of the agent's reasoning and code generation steps.

- Proven Versatility: Capable of generating a wide range of applications, from basic HTML "Hello World" files to functional tools like calculators and countdown timers.

- Developed a Telegram bot providing automated answers about game rules, contests, and mechanics.

- Implemented intent-based NLP to accurately respond to player queries.

- Integrated with backend services to fetch real-time game and player information.

- Created and published a course on Unreal Engine 5 XR Mastery.

- Designed hands-on lessons covering Blueprints, gameplay logic, and real-time interaction systems.

- Developed practical demos to explain Unreal Engine workflows for beginner to intermediate developers.