From competitive informatics to competitive robotics

I entered Tudor Vianu National High School of Computer Science in Bucharest in September 2011 and enrolled in an informatics-intensive class: seven hours of C and C++ every week, for four years.

I loved it, and it is worth being precise about what I loved, because it was not the language. It was that a problem with an exact answer could be attacked from several directions, and that some of those directions were beautiful and most of them were not. That is where I found out I enjoy computational puzzles, which is a narrower and more useful thing to know about yourself than enjoying computers.

The ninth grade was purely that. I competed in the national Informatics Olympiad and did not reach the podium. I keep that sentence in because the interesting part is what happened next rather than how far I got: the school robotics team took me on in the tenth grade anyway.

The tenth grade: a robot, from a fixed kit

The robotics team competed with Lego Mindstorms in FIRST Lego League, and we won the national championship at the first attempt.

What I remember more clearly than the win is the shape of the problem. FIRST Lego League does not let you fabricate parts. You get a defined set of components and you solve the task with those and nothing else. That is a real engineering discipline and it is a genuinely good way to learn: when the parts are fixed, the only thing you can be clever about is how you combine them.

It is also, it turned out, not the thing I was looking for.

The eleventh grade onward: the constraint comes off

From the eleventh grade I moved to FIRST Tech Challenge, and stayed with it through the twelfth and into my first year of university. FTC does let you fabricate. We designed in SolidWorks and printed our own parts.

That difference is the whole story of this page. In Lego League the solution space was a box of components somebody else had chosen. In FTC it was effectively unbounded: if you could imagine a mechanism, you could draw it, print it overnight and bolt it on in the morning. The limit stopped being the kit and became what you could think of.

And that is where the thing happened that I cannot really convey to anyone who has only written software. I designed a component in CAD, printed it, and then the code I wrote moved that physical object. The loop closes in the real world, where tolerances exist, where a bearing binds, where the thing you simulated correctly still does not work because you were wrong about friction.

I went from enthusiastic about informatics to in love with robotics, and I have not really moved since.

Autovortex #5843

FTC was a step up in every dimension: real materials, real machining, a real drivetrain, and a design cycle measured in months rather than an afternoon. Code that had to survive contact with a match rather than a test bench.

Over those seasons we competed in and won national championships in South Korea, Russia, the Netherlands, the Czech Republic and Germany. Winning a national championship is what qualifies you for the World Championship, so those trips are why I twice ended up in St. Louis, Missouri, in early May of 2015 and again in May 2016, with the other championships falling one or two months ahead of it in each season.

The best season was the one after I left

The last of those seasons ran while I was already in Germany, having started my Bachelor in Hanover. I stayed on the team anyway, and it turned out to be our best: Autovortex #5843 reached its highest placing, fourth overall in the Franklin Division.

I have thought about why the strongest season was the one I was least physically present for, and I do not think it is a coincidence so much as a consequence. Being remote forces everything through a written channel. Decisions that had previously been made by three people standing around a robot had to be argued in text, which meant they had to be reasoned rather than felt.

That is the first time I encountered something I now do for a living: when a team cannot rely on everyone having seen the same thing, the specification stops being bureaucracy and starts being the only thing holding the work together.

Why Germany, and why Hanover

The move itself was a straightforward calculation. Germany offered stronger universities than Romania for what I wanted to do, and materially better prospects both during the studies and after them, in an industry that is actually here.

Hanover specifically was not a calculation at all. I have an aunt and an uncle there who were generous enough to take me in for the duration of my Bachelor, and a nineteen year old moving countries alone takes that kind of offer. The university came with the family, rather than the other way round.