Vision Researcher – AI & Multimodal Models

About the role

You will work across the full lifecycle of vision-language model (VLM) development—including data collection, training, evaluation, and production. The team’s scope will evolve as the field advances, and we seek researchers who can lead and adapt.

Key responsibilities

• Research vision-language architectures such as encoders, fusion mechanisms, pre‑training objectives, and scaling behavior.
• Design training methods (pre‑training, SFT, RLHF, DPO) tailored for multilingual VLMs.
• Investigate data strategies, including mixture designs, quality signals, and synthetic data approaches.
• Build evaluation frameworks and benchmarks, especially for Indic multimodal tasks.
• Study model failure modes, robustness, and interpretability.
• Collaborate with engineers to prototype ideas quickly and validate them at scale.
• Engage with the broader research community through open‑source contributions and collaborations.

Required profile

• Deep understanding of vision-language models, including training dynamics, architecture trade‑offs, and failure modes.
• Proven research track record demonstrated by publications, technical reports, or impactful shipped work.
• Rigorous experimental design skills to isolate variables and draw defensible conclusions.
• Strong PyTorch expertise for end‑to‑end experiment execution.
• Intellectual breadth to work across data, training, and evaluation challenges.