NEW Data & ML Engineer – eMobility (SaaS)

Deftpower; Arnhem; Hybrid; EUR 60,000 - 90,000; Full time

Benefits offered by this job

Holiday allowance

25 vacation days

Pension

Training budget

Laptop provided

Hybrid office in Arnhem

Job summary

Deftpower is building a flexible virtual power plant that aggregates EV charging sessions for energy markets. You will design data pipelines and develop prediction models to forecast charging behavior and available flexibility, delivering production-quality solutions.

You will work with AI engineers and a team of C#/.NET developers in a hybrid setting, with an office presence in Arnhem. English is used daily; Dutch is a plus, and you’ll move fast in a dynamic energy tech environment.

Qualifications

• 3+ years of experience building data pipelines on high-volume data and production-ready prediction models.
• Rigorous with data: test assumptions and quantify how far a prediction can be trusted.
• You take initiative and investigate questions nobody has asked yet.
• You turn data insights into clear decisions for technical and business colleagues.
• You work in English every day; Dutch is a plus.

Responsibilities

• Design and build the data pipelines behind predictions, from raw data to model-ready datasets.
• Keep data reliable: monitor quality and catch issues before they reach a model.
• Develop prediction models and put them into production when they outperform current solutions.
• Collaborate with .NET developers to integrate models into the platform and monitor live models.

Skills

Python

ML stack

SQL

Cloud deployment

Git & PRs

Data pipelines

Tools

Azure

Bicep

Job description

Find new ways to make EV charging cheaper and smarter, from the data pipeline to the model.

About Deftpower

We make EV charging less annoying. Drivers find a charger, pay with one app or card across Europe, and charge at moments that are easier on the grid.

You won't see our name on that app. You'll see ANWB's, or one of the other brands we power. We like it that way: we'd rather make it work than make it loud.

Three e-mobility experts founded Deftpower, and they still run it. You can walk up to any of them, or to anyone else, with a question. Most people tell us they learn more from their colleagues here than from any course.

We're 120+ people now, with the energy of a team of 30. We'll keep growing, just more deliberately than before.

About the job

The team. The Flexibility team builds a virtual power plant: we aggregate thousands of EV charging sessions into flexibility for the energy markets. That makes charging cheaper for drivers and helps balance the grid. You join two AI engineers and a group of C#/.NET backend developers, led by Gerben.

Why now. Our platform works and has grown rapidly. We’re now entering the flexibility markets, and that raises the bar for our predictions and data. This is the moment to step in and help build what those markets demand.

What you do. Prediction is at the core of this role, on two levels. For each driver and charging session, we predict behaviour: when the car leaves, how fast it charges and how much energy it needs. Across thousands of sessions combined, we predict how much flexibility the pool can reliably deliver to the market. Those predictions are only as good as the data and models behind them, and you build both.

• Design and build the data pipelines behind our predictions, from raw charging and market data to model-ready datasets

• Keep the data reliable: monitor quality, and catch issues before they reach a model

• Develop prediction models and put them into production as soon as they beat what’s running now

• Work with our .NET developers to integrate models into the platform, and monitor them once they’re live

  How we work. We work with Shape Up instead of Scrum: five-week cycles with a clear goal, and the team decides how to get there. You get a lot of freedom in how you solve the problems we face, and you own the result.
  

  For you

  You get trust from day one. Contracts, roadmaps and numbers are open to you. You make the call, and you own it.

  You learn fast here. AI is part of every team's work, and feedback goes in every direction. Every quarter, we celebrate together at D-Day, the company day we work towards.

  It's not always tidy. We grew fast, and not everything is figured out yet. If that sounds like an opportunity, you'll fit right in.

  • 8% holiday allowance, 25 vacation days and a good pension

  • €2,000 training budget a year and your own choice of laptop

  • 40 hours a week, or 36 if you prefer

  • On average 2 days a week at the office in Arnhem
    

    About you

    • You have 3+ years of experience building data pipelines on high-volume data, and you’ve put prediction models into production.

    • You’re rigorous with data: you test your assumptions and know how far a prediction can be trusted.

    • You take initiative and dig into questions nobody has asked yet.

    • You turn what the data shows into clear choices for technical and business colleagues.

    • You work in English every day. Dutch is nice to have.

      Must-have: Python and its data and ML stack, SQL, deploying models to the cloud, Git and pull requests, and curiosity about AI coding tools.

      Nice to have: probabilistic forecasting, Azure and Bicep, and experience in energy markets, EV charging or battery optimisation.
      

    • This is for you if…

      ... you see a gap and plug it, without anyone having to ask.

      ... you say it when you disagree, and then you commit.

      ... change gives you energy. When plans shift, you shift with them.

      ... you share what you know and make your work visible.

    • This is not for you if…

      ... you need a fixed plan before you start. Our plans change, and often.

      ... you wait for instructions.

      ... direct feedback feels personal to you.

      ... you want to work fully remote.