Description
Sarvam is hiring an open-ended AI researcher to investigate architecture, optimization, scaling behavior, training stability, and post-training methods for its next-generation foundational models. The role involves designing and executing large-scale ablations, running end-to-end pre-training experiments, translating findings into production-ready training proposals, and collaborating with infrastructure and data teams. Candidates need a PhD in machine learning, computer science, or a closely related field, at least three years of post-PhD research experience, first-author publications at top-tier venues, hands-on transformer pre-training experience, PyTorch and distributed-training expertise, and a strong experimental-design mindset.
