Arvid - Data Science Manager
Hey I'm Arvid, the Data Science Manager at Tibber👋
My journey here started about nearly nine years ago, when I wrote my master's thesis with the company, back when Tibber was much smaller than today. I stuck around, and since then I've worked across pretty much every corner of data science we do, from Trading to App Experience where I helped turn our data into insights that are genuinely useful for customers day to day. These days I lead the team driving that work.
Why Tibber - what made you join, and what's kept you here?
In short, our mission: helping homes become more energy efficient and helping the wider energy system run on more renewables. Our VPP is a big piece of that, turning batteries and EVs sitting in people's homes into something that genuinely helps balance the grid. That mission hasn't changed since I joined, and it's exactly what's kept me here for almost nine years now. But if I'm honest, it's just as much the people: fun, engaged, smart people who genuinely enjoy solving an extremely important problem together. That combination is hard to find, and hard to leave.
What does data science work at Tibber look like - what kinds of problems are you solving?
It's a real mix, and that's part of what makes it fun. Right now some of the problems we're deep in include: forecasting a home's load so we can improve optimization for customers with batteries and EVs; flexibility forecasting, so we know how much our VPP can actually trade with; clustering homes so we can give customers a meaningful benchmark for how much their home should use; and quantifying how much our optimization and VPP are actually worth to a customer.
Different problems, but they all come back to the same thing: using data to make energy smarter and more valuable for the people using it.
You're building an ML platform at Tibber - what's the vision behind it?
Historically, data science at Tibber has been spread across the organization, with teams solving problems close to their own domain. That's given us speed, but it's also meant some reinventing of the wheel when similar problems solved are slightly differently in different corners of the company. What we're building now is a common platform: a shared tech stack and framework so new DS products get built on the same foundations instead of starting from scratch every time. The goal is less duplicated effort, a much better developer experience for our data scientists, and ultimately faster, more consistent value for our customers.
How has AI changed the way your team works, and what does that look like in practice?
It's significantly shifting how we work. We've incorporated AI into a lot of our development, with agents doing things like reviewing pull requests alongside the team. We've also made changes to the way we structure our DS products for AI, breaking them into smaller, dedicated repos that make it much easier for AI to work alongside us.
There's a lot more ahead of us here, and we've got big plans for how we rethink the discipline of data science to thrive in an AI age.
What kind of person thrives in your team?
The kind of person who has fun doing data science! We work on important problems but we believe that having fun and doing good work go hand in hand. People who genuinely love collaborating rather than working things out alone tend to well at Tibber. It also really helps to be good at translating business needs into actual data science product. The best people on my team can sit with a fuzzy business problem and figure out what to actually build. If you'd rather build something great with people than build it by yourself, you'll fit right in.
What gives you energy outside of work?
I'm a father of two small kids, so free time is a bit of a luxury these days! When I do get some, you'll usually find me in the mountains - I've always loved hiking and skiing. I also love sneaking in a round of golf when I can. But honestly, right now the best part of my day is just watching my kids explore the world.
