Summary
Soham Ghosh is an AI Research Scientist at Mistral AI in Sunnyvale, combining a decade of industrial and academic research to advance visual understanding in AI. His work centers on multimodal machine learning and reinforcement learning, with current focus on training capable vision-language models and self-supervised representations. Before Mistral, he contributed to Google’s vision-language initiatives (MUM) as a Software Engineer and later Senior Software Engineer, and he conducted SLAM and meta-RL research at CMU under Ruslan Salakhutdinov. His career spans founding roles and internships across Asia Risk Transfer Solutions, Quettra, IBM Research Singapore, Alexa Smart Home ML, and Amazon, reflecting a hands-on ability to ship production-ready ML systems and data-driven products. He holds a BEng from Nanyang Technological University and an MS in Computational Data Science from CMU, with an exchange stint at UC Berkeley, underscoring a strong academic foundation and a track record of cross-disciplinary collaboration.
10 years of coding experience
8 years of employment as a software developer
Exchange Program, Computer Science, Exchange Program, Computer Science at University of California, Berkeley
Bachelor of Engineering (BEng), Renaissance Engineering Programme, CGPA 4.76/5.0, Bachelor of Engineering (BEng), Renaissance Engineering Programme, CGPA 4.76/5.0 at Nanyang Technological University
GCE 'A' Level, AAAAB/Distinction, GCE 'A' Level, AAAAB/Distinction at Raffles Institution
Master of Science (M.S.), Computational Data Science, CGPA 4.03, Master of Science (M.S.), Computational Data Science, CGPA 4.03 at Carnegie Mellon University
English, Bengali, Hindi