Harshit Sharma

Final-year AI/ML Engineer specializing in Computer Vision, RAG Systems, LLM Engineering, and AI Agent Development. Hands-on experience building real-time AI surveillance platforms, hybrid retrieval pipelines, multi-agent systems, and open-vocabulary vision models. Deep expertise in LLMs, Prompt Engineering, Vector Search, Docker, Kubernetes, and MLOps.

Microsoft Hackathon

Finalist with proven ability to ship production-grade AI systems end-to-end.

AI Developer Intern - Codeacious Technologies - On-site

(2025-12 - 2026-05)

• Designed and deployed a real-time AI surveillance platform with open-vocabulary anomaly detection over live video streams, removing manual review — threats defined in natural language with zero model retraining.
• Engineered a production-grade RAG pipeline (FAISS dense + sparse keyword retrieval, cross-encoder reranking, semantic chunking) for accurate, source-grounded LLM responses over large document collections.
• Built a multi-LLM agent desktop app (NemoClaw fork, TypeScript/Electron) with sandboxed agents and Claude API integration, connected to Slack, Notion, and GitHub via prompt-engineered workflows.
• Containerized full AI stack (Docker/Kubernetes), automated alert escalation via n8n, benchmarked YOLOv8/YOLO-World/Grounding DINO, and delivered NestJS/MongoDB + React monitoring dashboard.

B.Tech - Information Technology - Gautam Buddha University (2026)