AI/ML Researcher – Multi-Modal (India)

The Mission:
While everyone else is busy fine-tuning LLMs to write better emails, you’ll be building the brain that understands the world through pixels, frequencies, and sensory streams. You aren't just processing video; you’re teaching machines to perceive knowledge from the chaos of reality. This is search and retrieval engineer for creating next machine learning foundation for digital and physical AI.
What you’ll do:
Decode the World: Develop SOTA embedding models that fuse video, audio, and multi-sensory data into a unified "knowledge perception" layer.
Beyond the Frame: Solve the "spatial and temporal gap"—helping AI understand not just what is happening in a video, but why and what happens next.
Ship the Future: We don't do research for the "sake of papers." You’ll be shipping products that define the next decade of physical intelligence.
Who you are:
Strong formal background in math and statistics, with a degree from IIT or equivalent preferred.
Product engineering background of 3+ years in a deep tech company or lab needed.
You dream in high-dimensional embeddings and view the world as a series of complex, interlocking signals.
You have a deep hands-on expertise in Computer Vision, Stats or Maths, or Multimodal Learning.
You’re comfortable working in the "grey area" where there are no good StackOverflow/chatGPT answers yet.
You think "Ground Truth" is a challenge, not a constant.
What you get:
Opportunity to work with a world-class technology team led by former IBM Research and product executives
Salary range: INR 40L - 60L / year
Strong equity in the company

Work arrangement
No