AI Research Engineer

What Novyte Materials does

We're building an AI-powered materials discovery platform — think of it as giving research scientists a copilot that accelerates how new materials are identified, tested, and brought to production.

Our stack combines PyTorch Bayesian models, AI agents, LLM Toolchains, a FastAPI backend, and a React Tailwind frontend.

We're backed by Theia Ventures, working with real enterprise customers, real scientists and shipping software that runs on sensitive on-prem GPU infrastructure.

What you'll do

Read and critically analyze recent AI/ML research papers.

Reproduce promising methods and establish strong baselines.

Form hypotheses, design experiments, run ablations, and analyze results.

Prototype novel architectures/algorithms and turn ideas into working implementations.

Build reliable evaluation and benchmarking pipelines.

Work with LLMs, multimodal models, agents, reasoning systems, retrieval/memory, and scientific AI depending on the research direction.

Fine-tune and experiment with models using SFT/RL/post-training techniques.

Run experiments efficiently across GPUs and distributed infrastructure.

Maintain reproducible research code, datasets, experiment tracking, and documentation.

Contribute toward research papers, technical reports, benchmarks and open-source releases.

A very strong fundamental in Mathematics, Probability and statistics and Deep learning is a must.

The ideal candidate is someone who enjoys understanding research papers, questioning the existing approaches and implementing ideas from scratch.