Comend was my “first” for many things, including hiring. When Albert and I raised our second cheque, our first order of business was bringing on talent to build product. Like anyone running a process for the first time, we talked to a lot of people and tried many strategies. It feels like these days everyone is either hiring or looking for work. I suppose there really are two types of people in the world.

This post isn’t about the actual pipeline/system we had in place to evaluate candidates, this was about how we sourced them.

I’m organizing this into 3 main categories:

The things that didn’t help: 0 practical leads

The things that helped: at least 1 valuable lead

The things that worked: we actually hired them

❌ The things that didn’t help

LinkedIn job posts, search, and conversation

Doing this for the first time, I thought that LinkedIn would be the likeliest way to find our next team member. among the over 1000 people that submitted their profiles to my job posting, only about 10 were even worth a second look. It’s possible we were unlucky, or more likely that I cast too wide of a net and should have tried a different approach, but in the end this ended being a waste of time and effort. If I were to use LinkedIn again I wouldn’t bother to make a job posting, I would just post about the job and link to a page with more details. Getting the news that you’re hiring in front of your followers and connections is far more valuable than using it to hire strangers.

Don’t even get me started on the amount of AI generated crap we received.

Looking at recent graduates

On paper, recent graduates seem like a great place to look. They are available, eager, and usually willing to take a chance on a small company (and probably for cheap). You also get the fantasy that you can find someone unusually sharp before everyone else notices. The problem is that for an early startup, potential is not the same thing as readiness.

What we needed was not just raw intelligence or enthusiasm. We needed people who could operate with ambiguity, make good product and technical judgment, and get useful work done without much structure around them. That is a hard profile to find in someone who has never worked full-time before. Most recent grads are still optimizing for the kinds of signals that school rewards: polish, responsiveness, theoretical knowledge, and being able to interview well. Those are nice traits, but they are not the same as being effective inside a chaotic startup.

There was also a practical issue. If someone is very strong right out of school, they often have cleaner, safer options than joining a small startup with limited brand, limited compensation, and a lot of uncertainty. So the candidate pool ends up in an awkward middle: people who are too inexperienced for what you need, or good enough that they have options you probably cannot beat. In real life, all of my most competent friends either ended up full time at a company they interned for, or landed a very strong job before they graduated.

I do not think hiring recent graduates is inherently a bad idea. I just think it was a bad fit for what we were trying to do at the time. For a company that has more structure, more management capacity, and room to train, it probably makes a lot more sense. Waterloo publishes PDFs with a list of graduates across all faculties after every convocation. If you’re looking for interns, and especially if you can take advantage of SR&ED, It’s a great choice. For us, it did not help. What’s funny is that prior to running this process (and starting Comend for that matter), Albert and I didn’t have much experience aside from internships between us. If I’m being honest, I wouldn’t have hired myself.

✅ The things that helped

Tracking layoffs

I visited layoffs.fyi and trueup.io every day to see if any companies had been added to the list. This was post-pandemic and most firms were running a tight ship, which meant layoffs weren’t that common. But a few that occurred recently at the time caught my eye. I found that BenchSci had done a round of layoffs, and folks from biotech firms like that were exactly the kind of people we were looking for. You can turn a list of companies into a list of names by searching LinkedIn for people announcing that they’re looking for a new job or take advantage of the “looking for work” profile header.

Does this violate "A players hire A players, B players hire C players”? Assuming a normal distribution of candidates, A players are rare and already working at a MANG company you likely can’t outspend. So yes, we likely won’t find A players this way. But I believe like Moneyball, lots of strong candidates are overlooked or passed on for reasons that have little to do with their performance. Sometimes it’s just the wrong environment, other times it’s a disconnect with what they’re working on. The investing advice of buying when everyone is selling might just apply here.

We found 1 strong candidate this way.

Finding top contributors on Github

Something I always do before doing anything is check if someone’s already done it. I’ve developed a habit of checking github whenever I’m looking for anything from react components to API keys. This time however, I made a search for projects that mentioned rare disease. While that particular search didn’t return much, it got me going down a rabbit hole of digging into the projects that I like most and seeing the active contributors.

While scrounging repos, I was surprised to see that some larger projects often had dedicated discussions around finding jobs that made use of that project’s tech. For example NextJs ran a monthly discussion for jobseekers. You can always run a search like “looking for work” and find thousands of discussions, issues, and users.

For engineers in particular, the Github hunting method is really effective; you can easily find relevant folks based on their past work, technological expertise, and easily see their level of experience/skill based on their contributions.

We found 2 strong candidates this way.

Referrals

One of the first things I looked up when we started hiring was how people actually get hired. The broad pattern was obvious: referrals punch way above their weight. They make up a small share of total applicants, but a much larger share of actual hires, and referred candidates tend to convert to interviews and offers at much higher rates than cold inbound. Recruiters also consistently rate referrals as one of the best sources of quality hires.

That made sense to me, because hiring is mostly an information asymmetry problem. As an employer, you are trying to infer reality from polished artifacts: résumés, personal websites, GitHub accounts, LinkedIn headlines, and interview performance. Referrals help because they add social verification. I know people who have flat-out lied about their titles on a résumé. That gets a lot harder when there are mutuals who know what you actually did. A referral does not guarantee quality, but it does make the candidate legible in a way a cold application often is not.

That cuts both ways though, and it is why hiring friends deserves real caution. We put one of my friends through the process, and they did well. But a conversation with a mutual friend (who had worked with them in the past) changed my mind about making the offer. That was a useful reminder that “I know them” is not the same thing as “I should work with them.” Friendship can distort judgment just as easily as it can improve trust.

Referrals also have a more practical advantage now: they partially dodge the AI application swamp. Application volume on LinkedIn has surged, with reports of roughly 11,000 applications per minute and materially higher application volume than just a few years ago. At the same time, candidates are using AI tools to mass-apply, tailor resumes, and manufacture polish at scale. In that environment, a referral is valuable partly because it brings back friction. Someone has to actually put their name behind you.

So for me, the value of referrals was never just that they were convenient. It was that they reduced uncertainty. In a hiring market full of noise, optimization, and synthetic polish, they were one of the few channels that still felt like signal.

🏆 The things that worked

Telling everyone we were looking to hire

Honestly, being in hiring mode in this economy almost feels like boasting. There’s a feeling of pride that comes with having the means to grow your team. Naturally Albert and I took every opportunity to tell people that we were looking for folks. I found myself turning into that guy that makes every conversation one that works to their benefit. There isn’t really a strategy here, it’s more an attitude and a sense of urgency. Hiring was about finding people, and we only know so many people between the two of us, so it only makes sense to get the word out to as many people as possible. You can’t keep hiring a secret. This ended up working in our favor.

At tech meetups, hack nights, demo nights, and founder events
In casual conversations with friends, especially ones in tech or adjacent industries
When catching up with former coworkers, classmates, or past collaborators
In conversations with other startup founders, operators, and early employees
At coffee chats with investors, advisors, or mentors
When someone mentioned layoffs, hiring freezes, or being unhappy at work
At coworking spaces, coffee shops, and other places where startup people naturally overlap
When meeting people from adjacent roles who might know engineers, designers, or operators looking for a move
Any time someone asked how the company was doing, what was new, or what the team needed next

One beautiful spring day, Albert was at working at coffee shop, when a group of girls took a seat at the table behind him. Eventually, they started talking about jobs, when one of the girls mentioned that she was having trouble finding a software engineering job now that her current company at the time was doing layoffs. According to Albert, he sat there listening to their conversation for a while before turning around and introducing himself. We put Megan through our process and she came out on top out. The rest is history.

Looking inside our userbase

Aline Lerner has a good blog about this. Users (especially the ones that use your product the most) already have a grasp of how your product works. Many might even have ideas of what to improve or change. They understand the problem better than most. If you’re lucky (like we were), mission alignment is already baked in. Especially when looking to fill a user-facing, for us in rare, it only made sense to bring someone from the community.

When we started our formal process, we had already launched our first product Librarey, which was receiving strong traffic on a daily basis. Due to the nature of the product being very specifically for rare parents, we ended up getting a lot of them visiting and using the site. This ended up being critical for our growth as a company. We met Betty because she was a rare mom, member of her newborn son’s disease patient group, and a user of the platform. Before that she ran product and marketing at several traveltech companies. When we learned that she was looking for work it felt like a perfect match. We started her off part time, and she eventually became a full time member of the team. We were lucky to have met her.

Rare disease touches people lives in unexpected ways. Often in an equal, yet unfair way. But this means that no matter who you’re looking for, there’s at least a 1/10 odds that the perfect candidate will have some connection to rare.

Conclusion

There were a lot of things we didn’t try. I might try them next time:

Working with a strong niche recruiter. Especially one who already knows your market, role shape, and stage. We actually consulted with a few recruiters before we started. This may be better suited for more senior hires.
Hiring contractors first, then converting the best ones. Lower-risk way to test execution, speed, and communication before making a full-time bet.
Running small events as a sourcing tool. Office hours, technical talks, domain roundtables, community webinars, founder breakfasts. Good people self-select into rooms like that.
We didn’t try poaching from other companies or competitors. Taboo? Maybe. But it seems like lots of people do it without shame.

In truth, we didn’t have a process. We tried many things in parallel and kept doing things that was working and stopped doing things that weren’t. It got easier once we stopped assuming the best candidates would come through the most obvious channels. The highest-signal sources weren’t polished job boards or endless inbound applications. They were people already close to the work: contributors, communities, referrals, users, and even chance conversations. That was the real lesson for me. The closer someone already is to the problem, the product, or the world, the better the odds that the fit will be real. If I ever had to do it again, I’d spend less time opening the funnel and far more time going where aligned people already are.