Nischal R Ellur

I’m a Computer Science and Engineering student who ships production-ready AI and backend systems. In my Campus Research Project at Unisys India, I deployed an air-gapped API automation platform that parses OpenAPI specs into executable workflow definitions and MCP service stubs, running with local LLM inference and zero cloud dependency.

I also built a privacy-preserving semantic search layer over live API documentation using ChromaDB and Qdrant, designed to prevent external data egress and integrate directly into the production search workflow. Previously, as a Project Intern at Honeywell, I delivered an end-to-end ML pipeline for live ground operations—modeling dynamic taxiway topology as a graph and serving conflict-minimized path predictions.

I enjoy turning hard real-world constraints into working systems. On my projects, I built a distributed GPU compute cluster with a unified OpenAI-compatible REST API gateway, created a real-time multiplayer platform with Docker and DigitalOcean deployment, and developed NeuralDSA—an adaptive DSA tutor with a FastAPI backend and a React knowledge graph frontend, backed by AWS Bedrock evaluation.