Endlytic

During the development of Endlytic, a collaborative project, I worked on building an intelligent platform that empowers developers to query Postman collections using natural language. The goal was to simplify navigation across large-scale API datasets by integrating a Retrieval-Augmented Generation (RAG) pipeline that delivers contextual and precise responses.

I contributed to designing a microservice-based architecture leveraging gRPC for high-speed communication between services and RabbitMQ workers for efficient task distribution. On the AI side, I implemented a RAG pipeline using LangChain, Pinecone, and Gemini embeddings to enable semantic search and context-aware LLM responses.

The frontend, built with Next.js, NextAuth, and Shadcn UI, provides a modern and intuitive interface, while the backend integrates with AWS S3 for secure storage and retrieval of Postman collections. The system was deployed using Vercel for the frontend and AWS EC2 for backend services, ensuring scalability and reliability.

This project significantly enhanced my understanding of distributed systems, LLM integration, and cloud infrastructure while reinforcing best practices in modular system design and microservice communication.

• Developed and orchestrated microservices using gRPC for seamless inter-service communication.
• Implemented RabbitMQ workers for efficient parsing and background job handling.
• Built a RAG pipeline with LangChain, Pinecone, and embeddings to enable intelligent context-based querying.
• Designed a secure and responsive Next.js frontend integrated with AWS S3 for data storage.
• Contributed to a collaborative development process, ensuring smooth integration between services and components.