Who we are 👩‍💻👨‍💻

The average cost of developing a new drug is over $1 billion. A crucial aspect that impedes such research is the unpredictability of chemistry - it takes lots of human effort to figure out how to make a potential drug candidate.

Our mission at Molecule.one is to solve the unpredictability of chemistry by leveraging AI. We believe that by combining our state-of-the-art machine learning models with public and proprietary chemistry data, we can ultimately develop an AI Chemist - an AI agent that can autonomously complete tasks done by a chemist.

We have developed the first commercially available chemical synthesis planning software entirely based on deep learning models. We were also the first start-up from Central and Eastern Europe to debut on the TechCrunch Disrupt SF stage in 2019. In 2021, we have closed a funding round of 4.6M USD supported by top US & European investors.

In August 2023, we announced our strategic collaboration with CAS - the owner of the world’s largest database of chemical reactions - to develop the world’s best platform for Synthetic Accessibility Scoring, which allows evaluating how hard it would be to make a molecule in the lab and significantly reduce the number of required experiments. We also run our own HTE (High-Throughput Experimentation) chemistry laboratory, which allows us to perform and analyze chemical reactions with cutting-edge throughput, providing customized data to improve our AI models.

Thanks to the above developments, we have access (in-licensed and growing self-generated) to the most comprehensive organic reaction dataset, and there are many exciting opportunities to build AI solutions on top of it.

We are a cross-domain team of people from different backgrounds, such as organic chemistry, software development, and machine learning. Currently, we have more than 20 people on board and counting! 💪 See the bottom of this page for more details on our core values.

What can you expect? 🧭

As a junior machine learning engineer, your responsibilities will include:

• Designing and implementing ML pipelines for predicting outcomes of chemical reactions, based on the largest reaction database available
• Developing our synthesis planning algorithm
• Following and applying relevant literature and research in ML
• Contributing regularly to our common codebase
• Collaborating closely with other teams on moving ML models to production
• Performing research on relevant topics. We regularly try to push the envelope of what is possible (including publishing twice SOTA results in reaction generation)

You will report to our Head of Machine Learning and work within our Machine Learning team. Your work will be interdisciplinary and also involve interacting with chemists who evaluate our solutions.

To understand better how we work at Molecule.one, please read the “Appendix: Our Values” section at the end of this document.

Your profile 👤

As an ideal candidate, you should have:

• A solid understanding of machine learning and deep learning foundations