Summary
Sukrit Chatterjee is an AI Research Scientist based in Bengaluru with four years of experience building and pre-training cutting-edge LLMs and multimodal models. He has led development of ALPHA 2.8, a 2.8B decoder model achieving SOTA on GSM8k, HumanEval and HellaSwag, and fine-tuned open-source models that outperform benchmarks like MMLU. Sukrit also worked on Veagle, a multimodal LLM that advances representation learning beyond BLIVA, Llava, and mPlugOwl, and has contributed to AI agent projects such as Kelly and Supercoder2. Comfortable spanning research and engineering, he moves models from pre-training to fine-tuning and production platforms, having shipped foundation-model work at SuperAGI and now at Atomicwork. A BTech graduate in EEE from Bennett University, he blends rigorous engineering with rapid iteration on GenAI research—often squeezing SOTA performance from compact model families.
4 years of coding experience
2 years of employment as a software developer
Bachelor of Technology - BTech, Electrical, Electronics and Communications Engineering, Bachelor of Technology - BTech, Electrical, Electronics and Communications Engineering at Bennett University