Read time 5 min
Our blog series takes a deep dive into the Reaktor way of working – giving you a behind-the-scenes view into the everyday joys, struggles, and wonders of working in a team.
Cast of characters
Eemil, machine learning engineer
Joni, data scientist
Mikko, data scientist
Client, Orion, started 03/2018
Picture the scene: three guys are sitting in a room on the fourth floor of Reaktor headquarters in Helsinki. It’s quiet, they’re working hard, headphones on and eyes on screen. On the wall, an army of post-it notes run riot, saying something cryptic like ”Raimo sorting topics.”
Eemil Väisänen, our jack of all trades, has worked at Reaktor for 2.5 years. Before that, he went through the ups and downs of start-up life as a developer, chief technology officer, and data engineer.
Joni Suorsa started his Ph.D. on theoretical physics and worked on artificial intelligence in a small start-up before joining Reaktor a few years back after being hit by a feeling of “been there done that” and thinking it was time to learn something new.
Mikko Koskinen finished his Ph.D. on physiology and neuroscience, after which he worked on his postdoc in Sweden and took the position of a data scientist in a hospital.
The trio are building a system with three services that Orion, a pharmaceuticals company, can use for exploring public information related to patents, clinical trials and publications. With the help of AI, it’s easier to further refine the information found in the articles and use the right data to support the right decision.
Let’s start from the beginning – what is a typical Reaktor project like?
Eemil: Good question. I think we have an internal narrative about the “average project”. Something we claim to be the norm.
Joni: I guess a mainstream Reaktorian project has a team sitting at the customer’s office together with the product owner and an assignment that is usually quite clearly defined. The case we’re currently working on doesn’t really tick any of those boxes. At Orion, R&D is done in several different places around the world, so we wouldn’t have face time with everyone anyway.
So you’re working on an AI project for a global pharmaceutical company?
Mikko: Yep. Orion basically asked us to study and enhance their research and development process and bring new AI based elements to it, which is quite an extensive job, at least from the pharma industry point of view. It’s exceptional in the sense that there’s no single problem we’re solving. Our main solutions in this project use artificial intelligence. We’re trying to introduce new working methods to an industry that’s quite set in its ways.
Joni: But it’s not like we’re trying to fit a square peg in a round hole. We have the relevant know-how as a team and we use artificial intelligence only when we see it as the best option.
Sounds like serious business.
Mikko: Medical documentation is something you can easily mess up. There’s good reasons for procedures to be strict and specific about what you can do.
Joni: I think we’ve succeeded because we started small. We’ve had a really limited scope.
Mikko: But the work was still quite obscure in the beginning. Even though I’m familiar with the research in the field, I felt dumbstruck listening to the seven-hour introduction to Orion’s medical research and development work. I didn’t know what to start with.
Eemil: The good thing about Mikko is that he can navigate between different layers of the business. He can do the strategy talk and jabber on about the high-level business needs and then come to us with very concrete ideas.
Let’s talk about you as a team. Do you feel like things work?
Eemil: We have a really good combo of expertise between us. Everyone knows something about everything, but we all have our own speciality.
Joni: I think it’s good. Sometimes there’s long periods when we only do infrastructure, which is not my core expertise, so I grumpily tap away.
Eemil: Yeah, I sometimes feel like we’re three cowboys doing our own thing under the same umbrella of a project.
Mikko: The best part is when we’re working intensively on the same stuff. Everybody knows the score and the focus is good.
Joni: And when we’re in the cowboy phase, we still know what the others are working on thanks to our daily stand ups, Kanban boards, and regular chit-chats. It’s not like “What in the world has Eemil been up to for the past three months?”
Eemil: I’m the kind of person who can start focusing on a detail whereas Joni moves faster. This is a lame comment but I really like working with him. I appreciate the fact that he likes to read so much.
Joni: The reading part doesn’t always lead to anything though.
Mikko: Eemil, on the other hand, is our stabilizing pillar. This was an IT project for two data scientists but I’m happy we got him along for the ride. He’s also a perfect fit for our dubious communications matrix.
Do you ever manage to get out for some fresh air?
Eemil: Yep. We’ve had dinners and beers.
Mikko: And we’ve gone to see a play at the theatre. That was fun.
Joni: Once we took the train to Turku and it got stuck overlooking a field for hours. We sat in a small cabin blabbering away with our former colleague Ferkku and some random student.
What’s your overall take on working at Reaktor?
Mikko: I really want to emphasise the importance of a good team. It’s the basis for everything.
Joni: And I appreciate the freedom and the opportunity to experiment, thanks to the fact that Orion really trusts what we do.
Mikko: It’s always good with a lot of data. Massive amounts of data.
Eemil: What I liked about start-ups is that you can do pretty much anything no matter what your title is. And we have that element at Reaktor, too. When I started, one of my friends said: “Ah, so you found the workplace that no one leaves.”