Senior MLOps Engineer
Zendar
Paris, France
Posted on May 27, 2025
Are you tired of your good old corporate job, working on one project for half a year? Is your passion for making great infrastructure being kept in a box labeled “for later”? Do you enjoy working with smart people and do you want to make their developer experience great?
Zendar is looking for an experienced Infrastructure Engineer who understands engineering needs and can work with engineers to make their life easier and more productive. Our team is a testament of the ability to unite engineers of different backgrounds over a common vision. We do not shy away from reality and its limitations but we also dream big.
You don’t have to fit in a box. We are not looking for a superhero. But we might be looking for you!
- Department
- Engineering
- Employment Type
- Full Time
- Location
- Paris - Île-de-France
- Workplace type
- Onsite
Your Role
To support our growing team, we’re looking for an MLOps Engineer to elevate our processes and technologies to the next level. If you’re excited about designing and managing the ML data lifecycle and optimizing ML model training and developer experience, we’d love to hear from you!
You will join our small team of Infrastructure engineers located in Berkeley, California and Paris, which means that not only will you have a lot of influence in shaping all of MLOps/DevOps/IT at Zendar, but you will also have a lot of responsibility for juggling many priorities and support tickets. This role is perfect for someone who is comfortable managing multiple responsibilities simultaneously, enjoys interacting frequently with other engineers, and is opinionated about what to prioritize and how to build things.
Key Responsibilities
- ML Data Management:
- Maintain and grow a robust and reliable data lake in GCP for our proprietary training and evaluation data.
- Grow our data lookup capabilities for data selection for labelling and model training.
- Scaling ML workflows:
- Take ownership of a high-capacity batch job system (Apache Airflow) for compute workloads.
- Design and optimize compute-heavy workflows for ML model training.
- Bring improvements to ML models release and version tracking.
- Support for Infrastructure:
- Be part of small Infrastructure team and serve as liaison to support Paris & German office Infrastructure (IT, CI/CD, VM maintenance, data upload issues)
- Setting and Maintaining Infrastructure Standards:
- Be a go-to resource for solving infrastructure-related challenges, fostering a culture of ownership and accountability in the team.
What We Look For
- Ability to work from the office in Paris at least on Monday, Tuesday, and Thursday. Wednesday and Friday are optional in-office days.
- 5+ years in ML Infrastructure engineering, MLOps or similar
- Deep experience with cloud computing platforms (GCP or other) and Cloud-native ML workflow.
- Broad knowledge of data engineering and MLOps tools (Airflow, Kubernetes, Terraform).
- Proficiency with Python and Bash.
- A proactive approach to solving problems and willingness to take on new challenges as they arise.
- Somebody with solid opinions based on experience, yet open-minded and flexible to find the best solution for the situation.
Bonus Points
- Prior experience of supporting ML research or data collection & management.
- Hands-on experience of driving the transformation from on-prem to in-cloud environment.
- Familiarity with build orchestration technologies and dependency management tools, comfort with containerization (Docker).
- Experience with CI/CD tools like GitHub Actions and a solid understanding of deployment workflows.
A track record of excelling in ambiguous or evolving roles.
Example of current projects to be involved in this role:
- Training ML models on data from our test vehicles benefits greatly from creating content-based subsets. Leveraging existing camera object detection model and automating data labeling is the logical next step, significantly reducing manual effort and streamlining the data pipeline. As the research requirements change, the camera model integration needs to seamlessly adjust as well.
- Our ML team uses the raw sensor data to train their models. Training a model consists of several processing steps, generating different versions of processed data. It would be beneficial to unify different ML workflows and enable storing these processed data to speed up the processing times.
- Zendar’s fast-paced environment puts strain on the speed of iterations during ML research and model training. With a deadline around the corner and models to be trained and retrained, the need to train fast becomes critical to our success. Choosing the right VM is just part of the solution as the right CPU & GPU utilization and script optimization can be more of a dealbreaker in most scenarios.
- Having large datasets at three geographical sites results in different challenges and some synchronization between locations is inevitable. We are currently working with different NAS devices, on-site compute platforms and localized cloud buckets but as we grow, our infrastructure needs to grow as well, in the most cost-optimal way.
What We Offer
- Opportunity to make an impact at a young, venture-backed company in an emerging market
- Collaboration with smart and motivated engineers and ability to execute your vision in a high impact role
- Competitive salary
- Lunch allowance, Carte Navigo and free snacks in the office
Zendar is committed to creating a diverse environment where talented people come to do their best work. We are proud to be an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, gender, gender identity or expression, sexual orientation, national origin, genetics, disability, age, or veteran status.
About Zendar
Zendar develops high-resolution radar imaging systems that enable next-generation safety features and driver-assistance technology that is reliable in all conditions and at sufficiently long range for high speed driving. Zendar has pioneered software-defined radar technology which fuses data streams from multiple sensors to build a more accurate scene and utilizes artificial intelligence models that interpret RF spectrum to understand the environment.
Zendar is headquartered in Berkeley, CA and has engineering offices in Germany and France.
Zendar has a diverse and dynamic team of electrical, mechanical, RF, algorithms and software engineers with a deep background in sensing technology. Backed by leading VCs and automotive Tier-1s, Zendar has raised more than $50M in funding and has established strong partnerships with industry leaders. Our team has more than doubled in number over the last year as our technology has gained traction.
Our Hiring Process
Stage 1:
Applied
Stage 2:
Review
Stage 3:
Hiring Manager Screen
Stage 4:
Technical Interview
Not quite right? Register your interest to be notified of any roles that come along that meet your criteria.