Novyte Materials is building a frontier AI engine for materials discovery, a system that learns the structure, physics and governing rules of matter, and uses that intelligence to propose entirely new materials for real-world synthesis.
We operate at the intersection of machine learning, scientific computing and materials physics , where datasets are sparse, search spaces are astronomical, and model decisions directly influence real laboratory experiments.
If you want to work on AI systems that reason about the physical world, not ad ranking, recommendation systems or standard NLP, this is the place.
Location : Mumbai / Bangalore
Role Description
We are hiring a Founding AI Engineer to design and build the core intelligence layer of Novyte’s platform. This is an on-site role in Mumbai / Bangalore, working closely with the founder.
You will architect generative, predictive and physics-aware ML models that operate across multimodal scientific data : graphs, crystal structures, text, spectra, simulations and high-dimensional descriptors. Your work will meaningfully shape the technical DNA of the company.
This is a founder-level role with end-to-end ownership. You will :
- Design next-gen ML architectures (GNNs, Transformers, Diffusion-style models)
- Build physics-aware representation learning systems
- Develop active learning and RL-based exploration loops
- Create scalable training pipelines and distributed training workflows
- Integrate ML models with first-principles simulation and scientific data
- Define foundational research practices, abstractions and model standards
- Work cross-functionally with computational chemists, platform engineers and research partners
This is not a routine AIML job. You will work in a frontier scientific domain with deep ambiguity, high complexity and massive real-world impact.
Qualifications
Required
Strong experience with PyTorch or JAXExpertise in graph learning, geometric deep learning, or multimodal MLExperience with RL, active learning or uncertainty-aware modelingStrong engineering fundamentals (clean abstractions, reproducibility, systematic debugging)Experience working with noisy, sparse or non-standard scientific datasetsAbility to read research papers, build prototypes quickly and reason from first principlesInterest in ML for scientific domains, not generic applied MLBonus
Experience with diffusion models, flow matching or generative modelingKnowledge of materials science, physics-informed ML or simulation-integrated MLFamiliarity with HPC, large-scale training or hybrid compute workflowsResearch publications or open-source contributionsPrior startup or founding experienceWhat Novyte Offers
Founding-level ownership and ability to define core IPA chance to build the intelligence core of a scientific discovery engineDeep, technical work with clear milestones and real-world validationA research-driven environment with full autonomy over modeling directionDirect impact, your models inform real lab experiments within months