Senior Machine Learning Engineer (all genders)

Babbel; Berlin; Remote; EUR 100.000 - 140.000; Vollzeit

Zusammenfassung

Babbel Labs is building the future of language learning with an AI-first approach, based in Berlin. We are seeking a senior individual-contributor to own real personalization subsystems end-to-end, from research to production, monitoring, and incident response.

You will work directly with the Principal Scientist to ship designs from specs, shape new features from the start, and make independent, confident decisions.

Qualifikationen

• Strong, hands-on ML engineering with real models in production.
• Experience with probabilistic modelling, latent-variable modelling, Bayesian inference or equivalent rigour.
• ML system evaluation: monitoring metrics, debugging, and safe rollout to live systems.
• Rigorous experimentation: offline benchmarks and online AB tests.
• Proficiency in TypeScript and Python, plus familiarity with AWS, Terraform, and CI/CD stacks.
• Coding agents are part of daily workflow and must be checked before use.

Aufgaben

• Own subsystems end to end from design to production and monitoring.
• Shape new features from inception, not just implementation.
• Take ownership of core personalization and mastery-tracking subsystems.
• Design rigorous evaluation plans including offline benchmarks and online experiments.
• Instrument and operate what you build, detect silent issues, and fix proactively.
• Deliver with coding agents and verify outputs before relying on them.

Kenntnisse

ML engineering

Probabilistic modelling

ML system evaluation

Experimentation

TypeScript/Python

Coding agents

Tools

AWS

Terraform

CI/CD

Jobbeschreibung

About Babbel Labs

Babbel Labs is building the future of language learning. We are an AI-first, independent company within the Babbel group, based in Berlin. Our teams bring AI, research and product together to ship experiences that set a new standard for how people learn.

The role

Babbel's learner-personalisation engine tracks what a learner has and hasn't mastered, and decides what they should practise next. We are actively pushing personalisation further: moving beyond describing a user and prescribing targeted practice, to adapting the learning journey ahead of them.

This is a hands-on senior individual-contributor role on the team behind it. You will own real subsystems end to end and make decisions about how to improve the personalisation engine. That includes everything from research and benchmarking through production, monitoring and the incidents that follow six months later.

How you'll make an impact

• Work directly with the Principal Scientist to take designs from spec into production, then keep improving what you've built on your own judgement rather than waiting to be told what's next.




• Help shape new features from the beginning, rather than only implementing a spec handed to you.




• Take real ownership of core personalisation and mastery-tracking subsystems, operate independently, and make decisions confidently and competently.




• Design the evaluation that tells you whether a change is real: a rigorous offline benchmark against a real baseline, and the online experiment that confirms or kills it.




• Run what you build. Instrument it, notice when it's silently wrong rather than only when it errors, and fix it before it becomes an incident.




• Deliver with coding agents as a matter of course, and verify what they produce before you rely on it.

Your skills and qualifications

• Strong, hands-on ML engineering that has shipped real models to production. Recommendation, ranking, scoring, or trust-and-safety systems under real user load are the closest match. Research or competition experience is a plus.




• Experience with probabilistic modelling, latent-variable modelling and Bayesian inference, or the equivalent rigour from an adjacent domain.




• ML system evaluation: monitoring metrics you define, debugging output that doesn't look right, and rolling out a change to a live scoring or ranking system without breaking it.




• Rigorous experimentation practice: benchmarking against a real baseline, running or reading A/B tests correctly, and the judgement to know when an offline improvement won't survive contact with production.




• TypeScript/Python as your primary languages, with enough command of our surrounding stack (AWS, Terraform, CI/CD) to ship and own your own service's delivery. This is not an infrastructure role, so depth there is not the bar.




• Coding agents are part of your daily workflow, and you check their output before you rely on it. You are neither dismissive of them nor careless with them.

Nice to have

• Experience with psychometric models, such as Item Response Theory.




• Graph ML experience at real scale, such as embeddings, graph neural networks, or relational modelling.




• Public technical work, such as open-source contributions, writing, or competitive ML.

Ways of working

This is a hands-on individual-contributor role on a small team that moves fast and works agent-first, but with a disciplined approach to delivery. It is remote and Berlin-friendly.

Diversity at Babbel

As part of our ongoing journey towards building a diverse, equitable, and inclusive company, we welcome everyone to apply, especially individuals who are underrepresented in tech. We are a learning company, inside and out, and we encourage you to apply even if you do not fit all the technical requirements — all candidates are assessed based on skills, qualifications, and our business needs. Please state your pronouns in your application, and let us know if you'd like to be addressed by a name other than the one appearing on your official documents. If you have a disability or special need, feel welcome to inform us so we can provide proper assistance in the application process.