Job opening: Machine Learning Developer

SCM is a growing and profitable scientific software company in Amsterdam, developing and marketing software in computational quantum chemistry and materials science for over 25 years. We have grown quite rapidly to around 20 developers over the last few years. To enhance our physics-based modeling with data-based modeling, we now aim to hire a scientific software developer with Machine Learning (ML) expertise.

Job description

Together with our current mathematics and ML developers, you will work on developing new functionality that applies state-of-the-art ML techniques to problems in computational chemistry. This new functionality will then be integrated into SCM’s Amsterdam Modeling Suite, which is used by thousands of industrial and academic end users around the world.

You will use ML primarily to increase the speed or accuracy of existing functionality. For example, to predict

  • molecular properties like dipole moments, polarizability, enthalpy of formation, chemical potential, solubility,
  • crystal properties like band structure, defect formation energies, or hardness,
  • atomic forces for large-scale molecular dynamics simulations (ML-based force fields), or QM/MM simulations
  • infrared, Raman, NMR, UV/VIS and XPS spectra
  • the initial density matrix for DFT calculations, or which points to sample during global or local optimization

You will need to

  • gather training data from databases or generate it yourself with active learning
  • find suitable descriptors or features of the molecule or chemical system
  • develop, choose, and train different ML models/architectures
  • estimate how reliable or accurate a predicted value is.

Our customers should also be able to train their own ML models.

ML methods in chemistry are developing rapidly. The position will require you to keep up to date with the latest developments, and judge how promising new methods from the academic world are.

You may also need to integrate open-source alternatives into our software. The code for the ML can be primarily written in Python, but most of our software is written in Fortran, so Fortran experience would be beneficial.

Job requirements

  • MSc or PhD in Machine Learning / AI / Data Science, OR
  • MSc or PhD in theoretical physics or theoretical chemistry, with excellent mathematics and programming skills
  • Interest in developing algorithms and software implementations to help our customers solve real-world problems in computational chemistry and physics
  • Excellent knowledge of and experience with Python
  • Proven hands-on experience in at least several of the following, preferably applied to chemistry or physics: deep learning, active learning, Bayesian techniques, Kernel-based methods, ML-based force fields, molecular and crystal descriptors, uncertainty estimation

Desirable knowledge and experience

  • TensorFlow and/or PyTorch.
  • High-performance computing infrastructure: e.g. job schedulers, optimized libraries, …
  • Fortran and/or C++ and modern software development practices in general.
  • Software development tools such as version control, debuggers, profilers, etc.
  • Density functional theory, force fields, and thermodynamics calculations
  • Generative models (e.g. GANs), reinforcement learning

We will primarily consider experienced applicants but may also consider candidates who are finishing their MSc studies soon, in case of suitable background and talents.

The working environment at SCM

SCM is a science-focused company which has been around for more than 25 years. The friendly working atmosphere at SCM is that of an informal small team of PhD-level researchers and scientific software developers doing intellectually challenging work. This environment provides much freedom and responsibility for shaping your own work. Originating as a spin-off company of the VU University, Amsterdam, SCM has its office space on the VU campus, at walking distance from the Amsterdam-Zuid train station. We also extensively collaborate with a growing number of academic groups and household-name industrial partners worldwide.

Thanks to its high quality of living and dynamism, Amsterdam has been elected as best European tech city and 4th most dynamic European city to work in, “best city for Millennials to live”, it consistently ranks high in Mercer’s annual Quality of Living surveys (11th best in the world, in 2019), and voted runner up in the best city survey by Time Out in 2021. The Netherlands is also number one on the English proficiency list of 100 countries.

What we offer

  • This position will initially be for one year but is intended to be extended and become permanent.
  • The position is available immediately but there is flexibility concerning the starting date.
  • Competitive salary related to experience, bonus, as well as relocation costs and secondary benefits (including ABP pension fund)
  • 34 holidays per year on top of national holidays.
  • Flexible working hours and the ability to occasionally work remotely.
  • This is a full-time (preferred) or part-time (at least 0.7 FTE) position based in Amsterdam, The Netherlands.
  • Tax benefits (30% of salary tax-free) are often applicable in The Netherlands (depending on circumstances) for a successful candidate from abroad.

Further information, applying

Detailed information on SCM and its software can be found at https://www.scm.com/

Those interested in this position are encouraged to contact us at jobml@scm.com for further information. Job applications can be sent by email to the same address until 30 September, 2021 (preferably sooner) or until the position has been filled. Applications should contain a CV, a cover letter explaining the motivation for applying and, if applicable, a list of software / ML projects worked on (with own contribution specified). References and source code may be requested at a later stage. The accompanying email should briefly stress already a few key reasons why you might be the most suitable person for the position.

Key concepts