At RESONIKS, we hire for more than skills, experience, or technical expertise. We look for people who genuinely live our values, people whose curiosity, collaboration, innovation, and drive for impact show up in what they do every day.

That's exactly why we're excited to shine the spotlight on Toni, our Machine Learning Lead.

Although Toni joined RESONIKS earlier this year, it somehow feels like he has been part of the team much longer. Quiet and thoughtful at first glance, he quickly surprises you with conversations that can effortlessly jump from machine learning to food, culture, woodworking, basketball statistics, or the occasional perfectly timed joke.

For Toni, solving hard problems is something he has been drawn to since childhood. A self-declared future researcher long before it became reality, he spent decades building expertise in sound, machine learning, and complex problem-solving. Today, he brings that same curiosity and passion into building intelligent systems that help machines better understand the world around us.

Beyond his work, Toni has been organising the DCASE Community Challenge since 2016 and continues to play an active role in this year's event. The challenge brings together researchers from academia and industry to tackle real-world audio and machine learning problems, with this year's emphasis on how AI can detect machine failures simply by listening, even in noisy factory environments.

We are incredibly happy to have Toni on the team and excited to share his story.

About Me

Hi, I’m Toni, Machine Learning Lead at RESONIKS, where I spend my days teaching machines to make sense of sound. In practice, that means working at the intersection of machine learning and audio: building systems that can hear, understand, and interpret the noisy, messy world around us.

Before RESONIKS, I spent nearly two decades at Tampere University researching sound event detection and acoustic scene analysis alongside some exceptionally smart people. Along the way, we published research, released open datasets that others actually found useful, and helped grow the DCASE challenge from a niche community effort into an international event that now attracts hundreds of researchers. That escalation still feels slightly surreal.

What keeps me hooked on this field is problem-solving. I enjoy hard problems, the kind that refuse to cooperate until you sit with them long enough, pull them apart, and eventually discover something that works. At RESONIKS, that means building systems that don’t just process sound, but actually understand what is happening in the world around them.

One thing I didn’t expect when joining RESONIKS was how quickly it felt like the right fit. The team is sharp, the hierarchy is refreshingly flat, and there is just enough terrible humour floating around to keep things human when the problems get difficult.

Outside of work, I play basketball, tinker with DIY electronics, build things out of wood, and occasionally disappear into side projects. Somewhere along the way, I also ended up creating a statistics platform for regional basketball leagues. Apparently, the data part of my brain doesn’t really switch off.

Why do you enjoy being part of the RESONIKS team?

It comes down to the people and the way we work together. The hierarchy is flat, collaboration feels natural, and the atmosphere is genuinely fun. The humour ranges from clever to deeply questionable, but underneath that, there’s a shared excitement about solving problems that actually matter, and you feel that energy in everyday work.

What gets you out of bed in the morning?

Unfinished problems. There is always something waiting to be figured out, and somehow the best ideas tend to arrive overnight, entirely on their own schedule.

How do our company values resonate with you?

Quite naturally, actually.

Innovation has followed me through most of my career, usually in the form of building tools and systems that did not exist yet, which is either passion or poor planning, depending on the day.

Diversity matters because great ideas rarely come from people who think exactly alike. RESONIKS is genuinely international, and mixing different backgrounds tends to produce better ideas, broader perspectives, and occasionally excellent bad jokes.

Impact is a big reason I moved from academia to industry. At some point, you want to see the thing you built leave the lab and do something meaningful in the real world.

What was the moment you knew you made the right decision to work here?

Pretty early on. I started mapping out the machine learning challenges ahead and realised two things at once: I had a strong sense of where to start, and the team around me complemented exactly the parts I didn’t have. That balance, expertise, trust, and room to innovate made it feel like the right environment very quickly.

What was your best day at work?

The days when the model works on real production data. Research datasets are clean, predictable, and polite. Production data is none of those things. So when something actually performs in the wild, that feels like a genuinely good day.

What’s the most exciting part of your job?

Honestly, its the problem-solving. There’s something deeply satisfying about wrestling with a difficult problem long enough that it finally starts making sense when the right approach clicks and suddenly all the pieces line up. That feeling never really gets old.

When you were a kid, what did you want to be when you grew up?

A researcher. I spent twenty years being exactly that, joined a startup, and somehow still spend a large part of my time researching. Some career pivots turn out to be less of a pivot than advertised.

What are your favourite hobbies or interests outside work?

I like tinkering as it conveniently covers a lot of ground. Sometimes it’s electronics, sometimes software, sometimes woodworking. The materials change, but the curiosity stays the same.

When do you come up with your best ideas?

Usually, when I’m nowhere near a computer, annoyingly. The best ideas tend to appear uninvited, sometimes during a walk, in the shower, while driving, and usually at the exact moment there is no practical way to write them down.