Summary
Pranith Karumanchi is a research-focused data scientist and engineer with ~5 years’ experience building and deploying LLMs and end-to-end ML systems, currently developing Generative Recourse Models at IIT Bombay. He has hands-on expertise in RLHF, SFT, LoRA and multimodal fusion, and has driven architecture and optimization changes that cut inference latency and boosted model performance by large margins in production settings. Pranith pairs research rigor with product delivery—having led LLM projects for startups and Turing that served at-scale applications—and mentors others, having guided 30+ projects and taught ML/RL/DL to 80+ students. He’s comfortable across the ML stack (PyTorch, TensorFlow, Hugging Face, Vertex AI, GCP) and brings a particular interest in reinforcement learning-inspired approaches (AlphaZero/AlphaGo) to generative and interpretable AI. A practical tinkerer as well as an educator, he focuses on making advanced models efficient, fair, and explainable for real-world impact.
5 years of coding experience
2 years of employment as a software developer
B.Tech Computer Science and Engineering, B.Tech Computer Science and Engineering at Vignan's Foundation for Science, Technology & Research
Master of Science - MS Data Science, Master of Science - MS Data Science at Vellore Institute of Technology
English, Telugu