Hiring is the most difficult but also the most important part of building a company imo

A few thoughts on this, and our philosophy around hiring:

1. We have a very high bar of quality when hiring a potential candidate for @SophontAI. It's much better to delay hiring for a position until you find the perfect candidate than hiring a subpar candidate asap. This may mean you may need to take on some additional responsibilities until someone for the role is hired, and this can be quite tough, but in the end it is often worth it. Because believe it or not, if you hire someone immediately just cuz you feel you're manpower-limited, if that person isn't the right fit for your company, the damage it brings to your company culture, roadmap, etc. will be much worse.

What you need is to be absolutely sure that hiring this person is going to make your company super successful and perfect for their job. So far I think everyone we've hired has made us feel this way.

2. What does "high quality" actually mean? This is quite a hard question, there isn't some checklist you can follow.

There is one trap people often fall into: just because someone has some degree (ex: PhD), graduated from some high status uni (ex: Stanford/MIT/etc.), or previously worked at a well-known company (ex: Google/Meta/etc.) doesn't mean they are high quality. They are often good candidates but not always will be the best fit for the position. If you over-index on these credentials, you will miss out on underappreciated talent. Two of our employees come from non-traditional backgrounds but one of them is an incredible researcher who thinks very deeply about our research problems while the other is a very smart, knowledgeable engineer, and we're very grateful to have them on board.

So what do you focus on? Focus on their accomplishments and tangible outputs. This could be papers, blog posts, github projects, etc. This could be actual contributions to projects you care about. This is a big part of our @MedARC_AI strategy btw.

Things that are important in a candidate will depend highly on the role you are hiring them for. For example, in a research-heavy role, traits like high agency, independence, good research taste, and so on. Nowadays, we also look for how candidates are using coding agents to accelerate their workflows.

By the way, a big thing to note about quality is that this is also defined with respect to a specific company. A high-quality candidate needs to have a good culture fit with your company. There are many really talented candidates we've talked to that just don't have the right culture fit but would do incredibly well in other companies. We especially have a bit of an idiosyncratic way of working (working completely remotely via Discord, leaning toward asynchronous communication, working with a broader community of volunteers, etc.) that may not be a fit for everyone.

With all of that said, I will say that a big part of deciding if a candidate is high quality is you'll know it when you see it.

3. How do you evaluate candidates? I think so many companies do this incorrectly. We've also done it incorrectly too, and honestly I'm still not sure we are doing it correctly but we're continuing to learn and iterate here. There's one key principle I believe in: try to evaluate people on tasks they will actually work on if they join. Think of it like AI benchmarks. If your benchmark task for hiring doesn't match the actual job you will hire a person to work on, then the benchmark task is useless!

This is why I think leetcode is frankly kinda useless. All that it will evaluate is if a person is good at completing leetcode problems, but that's not the job you're hiring them to do! Many people who would succeed in a given role might not be good at leetcode and vice versa.

This is why we instead try to do stuff like work trials, asking them to work on relevant research problems for a short period of time. And of course we don't need them to solve these problems immediately, but rather we evaluate how they tackle these problems, the way they think about the research process, how they communicate their results, etc.

Now I want to end this by saying: we're still learning how to hire. We're still a pretty small team and as we continue to grow we'll continue to learn. I would love to hear thoughts from others with more experience in hiring so we can improve our own approaches here :)