Researcher, Vision
Bengaluru • FullTime
Posted 4mo ago
Remote Work Policy
On-site
Employment Type
FullTime
Categories
AI Research Engineer
About the job
Sarvam is building the bedrock of Sovereign AI for India, developing a full-stack sovereign AI platform across research, models, infrastructure, and applications. The company partners with leading enterprises and public institutions, backed by prominent venture capital firms and collaborating with India's top brands. This role involves working across the full lifecycle of vision-language model (VLM) development, including data, training, evaluation, and production. We seek researchers who are comfortable with the evolving nature of the field and can take a leading role.
Responsibilities
- Research vision-language architectures, including encoders, fusion mechanisms, pretraining objectives, and scaling behavior.
- Design training methods (pretraining, SFT, RLHF, DPO) tailored for multilingual VLMs.
- Investigate data strategies, focusing on mixtures, quality signals, and synthetic data approaches.
- Build evaluation frameworks and benchmarks, particularly for Indic multimodal tasks.
- Study model failure modes, robustness, and interpretability.
- Collaborate closely with engineers to ensure ideas are testable at scale, prototyping rapidly and validating thoroughly.
- Engage with the research community through open-source contributions and collaborations.
Requirements
- Deep understanding of vision-language models, including training dynamics, architecture tradeoffs, and failure modes.
- Demonstrated track record of impactful research through publications, technical reports, or shipped work.
- Proficiency in rigorous experimental design to isolate variables and draw defensible conclusions.
- Strong PyTorch skills for end-to-end experiment execution.
- Intellectual range and willingness to work across data, training, and evaluation challenges.