Senior MLOps Engineer (m/w/d)
At a glance
Senior ML Platform & Ops role at Flexa. München · full-time · €95,000–€115,000 base.
aimlroles summary, based on the employer's posting.
What you'll do
- Create the data and feature flows behind forecasting and trading models operating across large distributed systems
- Use Flexa’s deployment and serving setup to launch models for both scheduled and live workloads
- Watch deployed models for drift, data quality, delays and failure patterns, then build visibility into those signals
- Bring engineering judgment to scientists’ model choices and checks, raising scale, maintenance and production concerns early
- Help set technical priorities and support junior engineers with design feedback, code reviews and coaching
What you bring
- A university degree in engineering, computer science or another listed analytical field
- At least five years in engineering, including substantial experience with ML systems running in production
- Strong Python ability and sound practices for testing, code quality, reviews, CI/CD and modular design
- Working MLOps knowledge across pipeline setup, deployment and monitoring
- Enough ML and statistics fluency to work through architecture tradeoffs with data scientists
Who this fits
This senior role suits an independent engineer who can shape their own tasks and collaborate across technical disciplines. It is office-first in Munich, requires C1-level English and is full-time with an indefinite term. You may also guide junior colleagues, though the role is not described as people management.
From the employer
Location: Munich, Germany
Employment Type: Full-Time, indefinite term
Salary Range: € 95,000 – 115,000 per year gross, depending on seniority level
Office First work setup
Language Requirement: C1 Level English
Act as the technical bridge between data science and software engineering, helping research models become reliable, maintainable production systems, and helping engineering understand what ML workloads actually need
Design and build the data and feature pipelines that support Flexa's forecasting and trading models at scale across hundreds of thousands of distributed systems
Leverage Flexa’s deployment, orchestration, and serving platform to bring models into production, for both batch and real-time workloads
Establish monitoring and observability for models in production, like drift, data quality, latency, and failure modes
Partner closely with data scientists on model design and validation, bringing an engineering perspective on scalability, maintainability, and production risk from early on
Champion engineering rigor and ML best practices to foster an open, data-driven engineering culture.
Contribute to the technical roadmap, anticipating scaling needs as data volume and model complexity grow
Opportunity to guide and develop more junior colleagues through design review, code review, and structured feedback
Be part of a cross-functional team of data scientists, software engineers, and other teams across Flexa's partner ecosystem
Mandatory Requirements
University degree in an engineering or analytical field (Computer Science, Mathematics, Physics, Statistics, Engineering or a related discipline)
5+ years of engineering experience, with significant time spent supporting or building ML systems in production
Proficiency in Python and software engineering best practices: testing, code quality, code review, CI/CD, monitoring, and modular code design
Solid working knowledge of MLOps practices: pipeline setup, deployment, monitoring
Enough fluency in ML/statistical modeling to collaborate effectively with data scientists and make sound architectural tradeoffs together
Independent, pragmatic problem-solving with strong attention to detail in a fast-paced environment
Excellent English communication and interpersonal skills
Cross-functional collaboration mindset across data scientists, software engineers, and partner-company stakeholders
Skills to Set You Apart
Experience in energy, power markets, or other near-real-time operational domains
Familiarity with orchestration tools (Airflow or similar), MLOps toolchains (MLflow, Sagemaker, or similar), and streaming systems (Kafka or similar)
Hands-on experience with large-scale data tooling: Spark, Dask, or comparable frameworks
Experience designing or owning near-real-time analytics and/or ML workflows, including observability
Track record of taking models from research into production on AWS or comparable cloud provider
This won’t be the right role for you if…
You don’t have the habit of defining your own tasks and have a preference for working in clearly separated functions
Virtual Share Options: we offer virtual share options to all our employees
Professional Development: annual development budget of €3,000 for coachings, trainings, books, and similar
Health & Sport Subsidy: company-subsidised sports facilities membership, or Public Transportation Subsidy
Lunch/Dinner Allowance Vouchers: allowance for meals on working days as digital meal vouchers
Work Equipment: MacBook or Windows laptop, iPhone (also for private use), and an ergonomic workplace setup with company-funded access to leading AI developer tools
Regular Team Events: knowledge sessions, afterwork, sports, offsites, Halloween, Pride Month, and more
Willi Richert, VP of Technology — flexa
Hi there!
I'm Willi, VP of Technology at flexa. I've spent the last 15 years building engineering teams around systems that have to make good decisions fast and at scale. Most recently at Lyft, where I led the mapping organization — 30+ engineers across five countries — and we moved more than 96% of rides off Google Maps onto our own mapping product. Before that I worked on machine learning and conversational AI at Microsoft Bing, and I did a PhD on learning in heterogeneous robot groups, which is a long way of saying that distributed decision-making has held my attention for a while.
What pulled me to flexa is that it's the same class of problem with something physical at the other end. We dispatch energy of the Enpal customer fleet in real time against energy markets. If our infrastructure is a few seconds late or a few percent off, customers lose money and the grid gets less flexibility than it could have had. That makes cloud engineering a first-order product concern here rather than a supporting function — which is why this role sits close to the decisions that actually matter.
How I work: I'd rather hand you the whole problem, context and constraints included, than a ticket. I care that engineers can see the consequences of what they build, and I'll be direct with feedback and expect the same back — the fastest way to lose a year is for everyone to stay polite about an architecture that isn't working.
You won't find everything already built. Some of it is greenfield, some of it needs replacing, and you'll have real influence over which is which. If that sounds like your kind of problem, I'd like to hear from you — and if you're not sure your profile is a perfect match, apply anyway and let's talk.
Looking forward to meeting you,
Willi