How we hire for infrastructure at Manychat: from first call to offer in 10 days
Most hiring drags on because teams have only a vague idea of who they’re actually looking for. We decided to fix that first.
The reason most hiring takes months has nothing to do with the number of stages. It’s that there’s no clear definition of the target. So companies compensate: they collect more CVs, add more interviews, compare candidates to each other — and still struggle to make a call.
We decided to remove that guesswork. Instead of growing the candidate pipeline, we built a process where a strong candidate can be identified and hired immediately.
I’m Dmitry, Head of Infrastructure at Manychat, and in this post I’ll walk through how we hire senior infrastructure engineers in ten days — by knowing exactly who we’re looking for and what we’re evaluating them against.
Know what you’re looking for
“Knowing who you’re looking for” isn’t about years of experience or a tech stack. It’s about understanding what this person will actually own. What problems will they solve? What level of autonomy is expected? What does good look like in six months?
Without answers to those questions, every candidate looks “somewhat relevant.” With them, the wrong ones are easy to spot.
Score candidates against the target, not against each other
When the previous point fails, the usual practice kicks in: companies build a large candidate pool and pick the best one from it. Candidates get compared to each other, not against a defined set of expectations. This creates delays — when a strong candidate comes in, you can’t commit immediately, because you need to see three to five more first.
We don’t do that. When the right candidate shows up, we say yes.
Filter hard at the top, move fast at the bottom
Knowing exactly who we want means we can filter aggressively at the top of the funnel. Here’s how it worked for our Senior SRE search.
We looked at two things on a resume. First, scale — large user bases, high-traffic systems, complex migrations, strict reliability requirements, or cloud/platform constraints. That experience can come from large companies, fast-growing startups, open-source platforms, or other high-scale environments.
Second, adaptability. The stack matters, but it’s not the whole picture. For a Senior SRE, we were looking for a generalist, not a specialist. Technologies change — what’s dominant today won’t be in two years. A CV that shows someone has switched technologies multiple times — started on bash, moved to Ansible, then picked up Kubernetes and AWS — is a reliable signal they can keep growing.
Out of 150 applications per hiring cycle, roughly 1,000 per quarter in total, around a dozen make it to the TA Screening Call. Five to seven reach the technical interview. Two get to the hiring manager interview. At least one gets the offer.
Hiring cycle
Clear evaluation criteria during interview stages
This is the key. Before any interview, we build not just question lists but scorecards: what good and bad answers look like. The scale is simple: needs improvement, meets expectations, exceeds expectations, and red flags. We place the candidate on it. That keeps evaluations consistent and removes cognitive load from the team. We write scorecards for both hard skills and soft skills.
For the technical interview, we have six topics: AWS, Kubernetes, Terraform, CI/CD, Observability, Security.
<a href="https://medium.com/media/84ee690a1e103580f0a685297bec3523/href">https://medium.com/media/84ee690a1e103580f0a685297bec3523/href</a><a href="https://medium.com/media/380a79fab054f48dbe6b89e6cedca914/href">https://medium.com/media/380a79fab054f48dbe6b89e6cedca914/href</a>
None of them includes live coding or similar exercises (they’re easily bypassed with AI), and we don’t see the point. We just talk. The candidate meets the whole team at once — we’re a small enough team so everyone joins the call.
Scorecards get filled out by each member of the team independently, on the day of the interview. If there are significant disagreements, we do a debrief — sometimes that means revisiting the criteria to arrive at a shared read on the candidate. Usually by the next day we know whether they’re moving forward. If the team says no, the process ends there. No wild cards from me.
The hiring manager interview works the same way — scorecards again. My job is to assess engineering maturity: what problems they’ve actually solved, what they’ve led, at what scale. I focus on four things: autonomy, data mindset, incident troubleshooting, project ownership.
To assess autonomy, I ask: Tell me about the last improvement you initiated without being asked. What was the problem, what did you do, and what was the result?
If there’s something to tell, I go deeper: Which metrics changed? Who pushed back? Specific dates, tools, results?
Here’s how the scoring looks for this dimension:
On our scale: “I upgraded Kubernetes” — needs improvement. “I migrated the whole company to a new stack” — meets expectations. “I led a cross-functional initiative — breaking up a monolith, moving between architectures, technically and organizationally complex” — exceeds expectations.
After this interview I know the candidate’s leadership level: what they can take on, how large a piece I can hand them. Sometimes the answer is the opposite — strong technically, but not ready for autonomous work, or missing the product mindset entirely.
The result of tech and leadership interviews is a filled-in candidate profile. For example, a candidate knows AWS well, has a gap in Kubernetes, but is solid on observability, and has led complex projects. From that I can tell whether they’re a fit for the team and whether they can carry what I need them to carry. The decision becomes easy.
What about the stages and timeline?
Application review takes up to two weeks but we push to move faster.
The hardest part logistically is getting the whole team together for the technical interview. Fitting the screening and the technical into one week rarely happens, but that’s what we aim for.
After the technical, I get on a call with the candidate within two days — already with the team’s feedback in hand. The final decision takes another day or two at most.
Total: two weeks on average from first call to offer
If this sounds like your kind of hiring process and you’re an experienced SRE engineer — follow me on Linkedin to stay in the loop on new openings. They are coming.
Top comments (0)