On the 19th of March, at the end of the work day, I grabbed my electric skateboard and did my last commute home through Marston Ferry Road as part of Mind Foundry; the next day would be my last at the company.

I worked there for almost 3 years as a senior research scientist in the field of quantum computing, specifically, in the calibration of quantum devices. As a physicist by training, this was a superb field to work on. I got to develop ML powered solutions, which I had been doing for over 5 years by then, to some of the problems of a field leveraging the most fascinating physics of our time. Not only this, I got to do all this with a group of brilliant and cool individuals that made each problem solving and catch-ups extremely fun and stimulating.

So why, you ask dear reader, would you quit a job like this? Let me elaborate on the spoiler in the title.

1-The Innovate UK quantum project had come to an end. This was a major milestone for our company’s work on quantum. As such, it provided me with the opportunity to pause, and ponder whether I wanted to continue with a new stage in the quantum work (if possible), move on to a new project at the company, or move on to a new chapter.

2-Robotics is cool¹.

Having decided that I wanted to go into robotics, I set out to spend some time before starting to apply for my next job to do mountain biking, learn some robotics, and a bit of German.

Oxford is known for many things, but not for its mountain biking. So I literally asked Claude “best place in germany for mountain biking and that has a good hackerspace” (to build a small robot). The answer was Freiburg. I had contemplated Freiburg in the past as a good place to live so it was settled. I arrived there on the first week of May.

As intended, I did lots of mountain biking, and built a robot at the local hackerspace. However, I didn’t put too much effort on my German. I barely was able to keep my Duolingo streak with freezes every other day.

Here is the so100 arm control demonstration

Also, I participated in the Le Robot Hackathon where we just trained a smolvla to place a cube into a square shaped hole, with push on the edge included if necessary.

Next steps

Now I’m working on exploring automated forward kinematics calibration and hopefully manage to train so100 with just one arm. If successful, this would half the starting cost.