Applied Science Works
Careers
We're looking for materials scientists, physicists, and chemists eager to shape the future of intelligent materials discovery.
Computational Chemist — DFT & Atomistic Simulation
About the Role
We are building a team at the frontier of atomistic simulation: combining first-principles calculations, molecular dynamics, and machine learning to model materials with quantum accuracy at scales classical methods can't reach alone. In this role, your center of gravity is the physics and chemistry: running density functional theory calculations, setting up and analyzing molecular dynamics simulations, and generating the high-quality quantum-mechanical data that everything downstream depends on.
We are hiring multiple team members for this group. This is a hands-on scientist position for a recent graduate. You will work alongside teammates focused on machine learning who train interatomic potentials on the data you produce — you do not need ML experience to excel here, though you'll have every opportunity to build it.
What You Will Do
- Perform DFT calculations (energies, forces, defect properties, migration barriers, phonons) using codes such as VASP, Quantum ESPRESSO, or CP2K
- Set up, run, and analyze molecular dynamics simulations in LAMMPS or GROMACS: structure building, force-field and protocol selection, equilibration, and trajectory analysis
- Design and curate DFT training datasets for machine-learned interatomic potentials, with attention to configuration diversity, convergence, and physical validity
- Validate ML-driven simulations against first-principles references and flag where models fail physically
- Automate workflows in Python (pymatgen, ASE) on HPC systems, working fluently with structure formats such as POSCAR and CIF
What We Are Looking For
Required
- M.Sc., M.Tech., or Ph.D. (recent or upcoming) in Chemistry, Materials Science, Physics, Chemical Engineering, or a related field
- Thesis, coursework, or internship experience with at least one first-principles code (VASP, Quantum ESPRESSO, CP2K, Gaussian) or one MD engine (LAMMPS, GROMACS)
- Python proficiency and comfort in a Linux / HPC environment
- Strong grounding in physical chemistry, statistical mechanics, or solid-state physics
Valued (we will train the right person)
- Experience with both DFT and MD, or with ab initio molecular dynamics
- pymatgen, ASE, or high-throughput calculation workflows
- Beyond-DFT methods (DFT+U, hybrids), NEB, or phonon calculations
- Visualization tools (OVITO, VESTA, CrystalMaker)
- Any exposure to machine learning applied to materials or chemistry
What You Will Gain
- Deep training in the data-generation side of machine-learned interatomic potentials — one of the fastest-growing specialties in computational science
- Mentorship from senior computational scientists and ML-focused peers to learn from
- Serious computational resources, a Python-first workflow, and a research-driven culture in Bengaluru
Compensation and Logistics
- Competitive salary commensurate with qualifications
- Competitive benefits: medical insurance, relocation support, conference travel
- Location: Bengaluru, Karnataka (on-site, 3/2 hybrid)
How to Apply
Send your CV, a brief note on a computational project you have worked on (what you did, what you learned, what you would do differently), and links to any code or publications to vivek.patil@appliedscienceworks.com with the subject line “Computational Chemist (DFT/Simulation) — [Your Name]”.
Machine Learning Engineer — Atomistic Simulation
About the Role
We are building a team that uses machine learning to transform molecular dynamics: training electronic-structure-aware interatomic potentials that deliver quantum accuracy at a fraction of the cost of first-principles methods. In this role, your center of gravity is the machine learning: model architectures, training pipelines, data-efficient learning, and the software engineering that makes it all reproducible and fast.
We are hiring multiple team members for this group. This is a hands-on engineering-and-research position open to anyone with the skills — from talented Bachelor's graduates to Master's and Ph.D. holders. You will work alongside computational chemists and physicists who generate the quantum-mechanical training data and validate the physics — you do not need a chemistry background to excel here, though you should be excited to learn the domain.
What You Will Do
- Train, fine-tune, and benchmark machine-learned interatomic potentials (e.g., MACE, NequIP, DeePMD, CHGNet) on DFT datasets
- Build and maintain training and inference pipelines in PyTorch and JAX (including Flax and JAX-MD), with clean, tested, well-documented code
- Develop active learning and uncertainty-quantification loops that decide which new quantum calculations are worth running — making the whole team's compute budget go further
- Integrate ML models with simulation engines (LAMMPS, GROMACS, JAX-MD) and materials tooling (pymatgen, ASE), including emerging AI-for-science frameworks such as SciLink
- Profile and optimize training and inference on GPUs; own experiment tracking, reproducibility, and model versioning
What We Are Looking For
Required
- Demonstrated experience building and training models in PyTorch or JAX — a substantial project, open-source contribution, internship, thesis, or publication. We evaluate what you have built, not the degree attached to it
- Strong Python engineering: you write code others can run, and you are comfortable with Git, Linux, and GPU environments
- Solid grasp of ML fundamentals: optimization, regularization, evaluation, and knowing when a model is lying to you
- A Bachelor's degree or higher in Computer Science, Applied Mathematics, Physics, Chemistry, Materials Science, Engineering, or a related field — B.Tech/B.E./B.Sc. graduates with strong portfolios are explicitly encouraged to apply
Valued (we will train the right person)
- Graph neural networks, equivariant architectures, or ML on molecules / materials / point clouds
- Active learning, Bayesian optimization, or uncertainty quantification
- Any exposure to atomistic simulation, DFT data, or cheminformatics (RDKit, pymatgen, ASE)
- HPC, distributed training, CUDA, or performance engineering
- MLOps: containerization, CI/CD, experiment tracking
What You Will Gain
- A rare specialization at the intersection of ML and physical science — ML interatomic potentials are among the most impactful and publishable applications of deep learning today
- Domain scientists as daily collaborators, so your models are grounded in real physics rather than benchmarks
- Serious GPU and HPC resources, a modern Python/JAX stack, and a research-driven culture in Bengaluru
Compensation and Logistics
- Competitive salary commensurate with qualifications
- Competitive benefits: medical insurance, relocation support, conference travel
- Location: Bengaluru, Karnataka (on-site, 3/2 hybrid)
How to Apply
Send your CV, a link to code you are proud of (GitHub or equivalent), and a brief note on an ML project you built (what you did, what broke, and what you would do differently) to vivek.patil@appliedscienceworks.com with the subject line “ML Engineer (Atomistic Simulation) — [Your Name]”.
Don't see a fit but think you'd be a good addition to the team? Get in touch — we'd still like to hear from you.