Shreyas

I'm a Deep Learning Engineer focused on building systems where models reason, act, and collaborate. With a foundation in physics and mechanical engineering, I specialize in applying AI especially vision + LLM systems to real-world tasks. I’ve built agentic AI workflows using LangChain, LangGraph, and ReAct, such as Linked Squad (a multi-agent LinkedIn optimizer) and DoqToq (a modular RAG system that gives documents a voice). On the vision side, I work with GANs, LaMa, UNet, Pix2Pix, Diffusers, and Fourier pipelines for inpainting, segmentation, and image generation. I focus on scalable, optimized deployment building pipelines with PyTorch DDP, Docker, and Flask and I enjoy turning research ideas into robust, usable systems. My toolkit includes PyTorch, Hugging Face, LangChain, LangGraph, MLflow, GCP, and more.