Computational Scientist

We're building a platform for molecular discovery that combines quantum chemistry, machine learning, and high-throughput automation. The foundation is already in production: GPU-accelerated DFT, semi-empirical methods (g-xTB), conformer search, ML synthesizability and property prediction, and generative molecular design, running across GPU, HPC, and cloud. We're looking for the person who takes it to the frontier.
*The role:*
You'll own the verification pipeline: the layer that proves a generated molecule is physically real, stable, and performs as predicted. Today it runs on DFT and QM validation. Your mandate is to push it into the capabilities we don't have yet: custom-trained ML interatomic potentials, molecular dynamics, and reaction kinetics. You'll write most of that code, set the scientific direction for validation, and harden what exists. Come build with us while the team is small. You'll have real ownership of your part of the platform, make the calls that matter, and see your work ship fast.
*What we need:*
- Demonstrated depth in at least one of: training MLIPs from first-principles data, molecular dynamics, or reaction kinetics. Working familiarity with the others, or a clear track record of picking up hard methods fast.
- Strong software engineering. Clean, typed, tested, modular Python. You can take research code to production and back, and own CI, Docker, and release hygiene.
- Solid quantum-chemistry fundamentals: DFT, semi-empirical / tight-binding (xTB family), and the statistical thermodynamics to compute properties correctly. You treat methods as tools you understand, not black boxes.
- Cheminformatics (RDKit) and honest ML, including leakage-free validation.
- Comfort running jobs on GPU/HPC.
- A self-directing, founder mindset. You scope your own work and move under ambiguity.
We care more about judgment, breadth, and the ability to learn than about checking every box. If you've taught yourself a method no one handed you, we want to talk.
*How to apply:*
Reach out with a short note and anything that shows what you've built (GitHub, thesis, a project you're proud of).
Sam Cogan [email protected] 052-758-6893