ML & Software Engineer (Junior/Mid) at Molecule.one

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.

Most recently, together with OpenAI, we demonstrated something we're very proud of: the first chemistry discovery where an AI model proposed a novel experiment that then held up in the lab. What made it work at that speed was automation, letting us run multiple experimental cycles in the time a study like this normally takes to run once. In December 2025, we also entered into a multi-year strategic agreement with W. R. Grace, a leading global supplier of specialty chemicals, to develop new, scalable ways to manufacture chemical compounds, starting with peptide building blocks.

This builds on years of work. We developed the first commercially available chemical synthesis planning software entirely based on deep learning models, and were the first start-up from Central and Eastern Europe to debut on the TechCrunch Disrupt SF stage in 2019. In 2021, we 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. In June 2025, our retrosynthesis platform was named the winner of the $1 million Chemical Innovation Challenge organized by Standard Industries.

We are currently developing the Maria platform, which unites AI, an automated microliter-scale HTE lab, and proprietary data into a single system. Guided by our data and AI models, the lab can produce molecules with far fewer failed attempts and at a fraction of the cost of conventional laboratories. The reactions it runs generate a unique dataset that we feed straight back into the model, closing the loop: the AI proposes experiments, the lab runs them cheaply and at scale, and the resulting data makes the AI better at proposing the next ones, giving us a system that keeps improving with every reaction it runs. Building that system is exactly what the Maria team does — the team we're hiring you to join.

We are a cross-domain team of people from different backgrounds, such as organic chemistry, software development, and machine learning. Currently, we have almost 30 people on board and counting! 💪 Read more about our core values on the careers page.

The Role 🧭

We're looking for an engineer who wants their code to make chemistry more predictable and push it into territory that isn't mapped yet. You'll work at the point where ML meets the physical lab, tightening the loop between our models and the lab itself: the faster experiments run and data flows back, the faster the models learn what chemistry can actually do.

This is a hands-on ML engineering role: you'll build models, but also the pipelines and automation that keep them fed with fresh, high-quality lab data.

Your responsibilities will include:

Designing and implementing ML and data pipelines for predicting outcomes of chemical reactions, based on the largest reaction database available
Building the automation layer that closes the experimental loop
Developing agentic approaches for proposing and prioritizing lab experiments
Building tools that accelerate chemists' work, improving the quality of the data we collect and ensuring we gather the data most useful for our models
Moving ML models to production
Following relevant ML literature and pushing the envelope where it counts (we've twice published SOTA results in reaction generation)

You'll report to our Head of Maria Platform and work within the Maria team. The work is interdisciplinary and involves close collaboration with the chemists who evaluate our solutions.

To understand better how we work, please read the values section on our careers page.

Your profile 👤

As an ideal candidate, you should have:

A solid ML/deep-learning foundation — you should be comfortable building, training, and reasoning about models
Strong software engineering skills — you write production code and care about the systems your models live in
You are a high agency person (https://www.highagency.com/). You are not scared to question authority, propose your own solutions, and work independently
Experience with building machine learning solutions with Python and PyTorch
A bias toward making things work end to end, you're drawn to the engineering that turns a model's output into a real result, not just the model in isolation
Problem identification, problem-solving, and critical thinking
Ability to work in a team in a fast-paced environment (including but not limited to agile methodologies)
Willingness and ability to quickly grasp new concepts
Communicative English (B2)

Nice to have:

Interest in chemistry 🧪 or biology 🧬.
Experience in research (for example: research thesis, blog posts, co-authorship of research publications) 📚.
Familiarity with JavaScript/TypeScript — we value people who can move across the stack when needed 🛠️
Knowledge MLOps/DevOps related technologies

What we offer 🎉

💵 Junior: 12000 - 16000 PLN + VAT
💵 Mid: 16000 - 20000 PLN + VAT
per month on a B2B or an equivalent on another form of employment
🕒 Full-time position with immediate start date.
🔭 An opportunity to work on novel and exciting problems.
💻 All the equipment you need.
⏱️ Flexible work hours and hybrid work policy (we have offices in Warsaw and Kraków)
📝 Opportunity to publish and present your work at top venues.
👥 A top team that will inspire you to be your best self.
💰 Equity package

💡 Interested? Please reach out to us via email: ai.recruitment@molecule.one by 10/07/2026. We can't wait to meet you and discuss the great things we can achieve together! Please do not forget to add to your resume the following clause: I agree to the processing of personal data provided in this document for realizing the recruitment process pursuant to the Personal Data Protection Act of 10 May 2018 (Journal of Laws 2018, item 1000) and in agreement with Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation).

The hiring process at Molecule.one ☕

↗️ Our hiring process consists of several stages. After each stage, we will either invite you to the next stage or discontinue the process. In the latter scenario, we will also give you feedback.

🇬🇧 The whole recruitment process is conducted in English.

0. CV Screening
As the first part of the process, we ask you to send us your CV. After examining your CV, we will invite you to the first interview.

1. Introduction interview
The first interview will last up to 30 minutes and take place online. During this interview, one of our team members will ask you a few questions about Machine Learning and software development.

2. Programming assignment
After the first interview, we will send you a short programming assignment. This assignment should take at most a couple of hours to complete. We will ask you to send us the completed assignment within four days of receiving it.

3. The final interview
After receiving and reviewing your programming assignment, we will invite you to the final interview. This interview can be conducted online or on-site in one of our offices, depending on your preference. The interview will last up to 90 minutes and involve some questions about your assignment solution, as well as more in-depth discussions about your experience in Machine Learning and software development.

✨ Interested? Please reach out to us via email: ai.recruitment@molecule.one
We can't wait to meet you and discuss the great things we can achieve together!