Ito

I currently have two roles. The first is managing the organisation and operations side of product development. Our product development brings together several technical teams — Nakazawa's web engineering team, Tobias's machine learning engineering team, and the machine learning research team — so I make sure each team's work runs smoothly. The second is planning the features we build into the product.

Nakazawa

Ito originally joined us as an outside contractor, and became an employee partway through, didn’t you? Our job on the web engineering team is to figure out the specs we need by talking things through with the product manager, then build them out — back end to front end.

Ito

The product manager Nakazawa just mentioned is also my role. I handle requirements definition — figuring out which clients miHub® should be solving problems for, what those problems actually are, and what features we need to build to solve them.

Tobias

As player-manager of the machine learning team, I work on improving the machine learning features miHub® provides. That machine learning helps clients get from their experimental data to the optimal answer as quickly and efficiently as possible. By the way, I also double as a site reliability engineer (SRE).

Nakazawa

Many of MI-6's web engineers, myself included, had no contact with MI before joining, but the more miHub® evolves, the more we're finding we need a deeper understanding of MI and our clients' work, too. At the moment we're running study sessions whenever we can fit them in, to help the team build up their understanding of MI.

Tobias

Nakazawa isn’t just an excellent engineer — it’s reassuring that they also lead the team through challenges like that. Running our machine learning platform depends on close cooperation with the product management team, the web team, the machine learning research team and the SRE team, so we communicate closely on Slack day to day and discuss things online whenever we need to.

Nakazawa

When I was job-hunting, what I really cared about was finding a specialised field where I could help solve social issues, while putting the mathematical modelling I'd studied at university to good use. That's how I came across MI, and I find it hugely rewarding to bring structure, through software development, to materials development — an extremely specialised and unstructured field.

Tobias

MI was uncharted territory for me too. But when I met the development team at my interview, I remember thinking, 'I could do good work with this team' — so I figured the domain knowledge was something I could pick up later.

Ito

For me too, being drawn to the people here was a big part of why I joined. Plenty of our team members want to contribute to R&D by driving MI adoption, and even our executives think not just about the company's growth, but about how to support the growth of each individual employee — I think it's a genuinely altruistic environment.

Nakazawa

Because we can talk things through constructively whenever we hit a problem, it doesn't take long to see a clear way forward, either. It's interesting how varied everyone's backgrounds are — I was surprised to learn, for instance, that Tobias majored in economics at graduate school.

Ito

Being part of this whole shift, as MI technology actually gets put to work in society, is incredibly valuable too. We're seeing more and more clients actually using miHub®, and that really feels like progress.

Tobias

It's become really clear that miHub® in particular has actually helped with a huge number of materials and product development projects. That's thanks partly to the product management team really digging in to understand the features and user experience clients need and designing around that, but at the same time, we on the machine learning team have been bedding in scrum processes while pushing development and operations forward at speed. I think that's what's showing in the results.

Ito

If I had to pick a memorable initiative, it would have to be the major update to miHub®. It was a challenge that tackled head-on something every research site felt was a problem, but that had been difficult to turn into a system. The web team had a huge amount on their plate, so this time Tobias’s machine learning team stepped in to help.

Tobias

Technically, we're probably no match for a professional web engineer, but we've been working closely with Ito to help the team get more done, faster. Ito's a product manager who thinks like an engineer, and makes sharp calls that bring the teams closer together.

Nakazawa

That really helped us. This update meant we had to improve the product and rethink the web team's structure at the same time, so it was doubly demanding. But thanks to that, we ended up talking to a lot more people outside our own team, and in the end we managed to evolve the whole development process too.

Tobias

Going through this update strengthened collaboration between web and machine learning and matured miHub® as a product, of course, but what really stands out is how much closer the teams got, making it easier for us to share the same goals.

Nakazawa

Our web team was still young, and we hadn't really found our structure yet, but Ito nailed down exactly what we were missing and gave us a long-term vision that covered not just the immediate work, but our people too. I think that guidance made a huge difference.

Ito

I think we're in a transitional period for MI right now, and at the same time a period of transformation for the company. I find it genuinely rewarding to help build a new culture in the world of R&D. The organisation needs to transform too, and it matters that we see ourselves as laying our own tracks rather than just walking along tracks someone else has laid — and that we enjoy doing it.

Ito

MI-6's strength is that high-calibre people make up a really strong team. A lot of our members come from research backgrounds, so we're quick to get to the heart of a problem, and on top of that, the development team's technical skills are strong too, letting us turn the PDCA cycle quickly.

Nakazawa

Everyone told me the MI study sessions I started within the web team were 'really interesting.' I think that's one of the best things about MI-6 — plenty of people are curious about fields outside their own specialism, even among us web engineers.

Tobias

Whatever your speciality, going forward it'll be important not to fixate solely on your own technical skills, but to think about the product with a long-term view.

Nakazawa

For MI to become something used naturally on the ground in research, we need to deliver UI/UX, data structures and usability that feel right for that setting, while also keeping the software itself robust and easy to maintain. I want to keep polishing miHub® more strategically, while building shared understanding across teams.

Tobias

For my part, my goal going forward is to make the value we deliver through machine learning more robust and sustainable. Even as the service gains more features and more people get involved, I want to keep validating that value without losing speed.

Ito

I find it really rewarding to be taking on this goal alongside my colleagues at MI-6 — building a future where using MI is simply the norm. I think we're in a transitional period right now, in terms of whether MI as a technology can really take hold in society. That brings plenty of challenges, but I get to have valuable experiences that only this moment makes possible. I want to keep enjoying the process of building a new culture, while gaining experiences that are only possible here.