Harsh Kasat

Developed multiple LLM agents using Hugging Face, Stable Diffusion, LlamaIndex, and Langchain, and implemented RAG models for context-free applications. Designed APIs with FastAPI for efficient model execution, utilizing Google Cloud Platform for authentication. Key technologies include Transformers and Google Cloud Platform.

Maintained frontend HTML, CSS, and JavaScript, managed email broadcasts, and handled server deployments. Developed scalable Django backend with Supabase and SQLite, implemented CI/CD pipelines with GitHub Actions, and optimized applications using Docker, load balancers, and Redis.

Implemented real-time car and people counting, and car number plate detection using OpenCV and deep learning. Developed scalable invoice-to-JSON conversion with large language models, created generative fills with Stable Diffusion and ControlNet, and architected microservices for production LLM models, optimizing performance to handle millions of requests.