Great read. As a hiring manager at Coinbase in 2021/2022, this is an accurate summary of the process. I wish notetaking AI was available back then, to help with consistency.
One of the bigger risks was the hire / no hire determination submitted in Greenhouse by individual interviewers. There was a reverse shadow process to calibrate interviewers when they started, but it wasn't an ongoing process.
When a company-level priority to 3x employee size in one year was introduced (called a Code Yellow), it made it difficult to maintain that consistently-high hiring bar.
Notetaking AI would have enabled more consistency & spot checking to ensure interview quality & alignment to the planned interview questions we were expected to ask. It would also give interviewers confidence they were within nominal ranges.
An internal AI audit process like the above is something I'd recommend all tech companies implement. It ensures consistency on interviewers and creates a better overall applicant experience.
A question about "Candidates and hiring managers would express their preference, and we landed candidates based on both preferences.". When the article say "we", are they referring to the HR department (vs Engineering Management)? I’m asking because there’s often competition between teams to secure the best candidates.
At Coinbase every opening (basically a PID that means you can hire someone) was globally stack ranked against all other PIDs in terms of urgency to hire.
You wanted to have your PID high on the list, otherwise you could go months without filling that role.
The process was eventually modified so that you'd automatically move up the global list if your role had been open for a longer time.
In practice, this meant an applicant could express interest in team X, but unless it was a referral from that team there was no guarantee that team X had a PID that was high on the global ranking list.
Team matching is like The Voice, you want multiple judges to turn their chairs, so you have some choice.
TL;DR Eng teams basically had a stack ranked list of openings that was tweaked every week. The higher your openings were on that global ranking, the more team matching chats you had.
My team was never super high on the list, so Eng Mgrs like me spent time to create slide decks to pitch the Eng that we had matching chats with. We needed to have a high conversion rate for anyone that matched with us.
Thanks for the detailed breakdown! I asked for more details because I suspected this might be the case, it’s a classic conflict that arises with this kind of cross-functional hiring setup. Still curious to know how exactly these PIDs were ranked and who was responsible for deciding those priorities ;-)
Great read. As a hiring manager at Coinbase in 2021/2022, this is an accurate summary of the process. I wish notetaking AI was available back then, to help with consistency.
One of the bigger risks was the hire / no hire determination submitted in Greenhouse by individual interviewers. There was a reverse shadow process to calibrate interviewers when they started, but it wasn't an ongoing process.
When a company-level priority to 3x employee size in one year was introduced (called a Code Yellow), it made it difficult to maintain that consistently-high hiring bar.
Notetaking AI would have enabled more consistency & spot checking to ensure interview quality & alignment to the planned interview questions we were expected to ask. It would also give interviewers confidence they were within nominal ranges.
An internal AI audit process like the above is something I'd recommend all tech companies implement. It ensures consistency on interviewers and creates a better overall applicant experience.
A question about "Candidates and hiring managers would express their preference, and we landed candidates based on both preferences.". When the article say "we", are they referring to the HR department (vs Engineering Management)? I’m asking because there’s often competition between teams to secure the best candidates.
At Coinbase every opening (basically a PID that means you can hire someone) was globally stack ranked against all other PIDs in terms of urgency to hire.
You wanted to have your PID high on the list, otherwise you could go months without filling that role.
The process was eventually modified so that you'd automatically move up the global list if your role had been open for a longer time.
In practice, this meant an applicant could express interest in team X, but unless it was a referral from that team there was no guarantee that team X had a PID that was high on the global ranking list.
Team matching is like The Voice, you want multiple judges to turn their chairs, so you have some choice.
TL;DR Eng teams basically had a stack ranked list of openings that was tweaked every week. The higher your openings were on that global ranking, the more team matching chats you had.
My team was never super high on the list, so Eng Mgrs like me spent time to create slide decks to pitch the Eng that we had matching chats with. We needed to have a high conversion rate for anyone that matched with us.
Thanks for the detailed breakdown! I asked for more details because I suspected this might be the case, it’s a classic conflict that arises with this kind of cross-functional hiring setup. Still curious to know how exactly these PIDs were ranked and who was responsible for deciding those priorities ;-)
Basically Director Thunderdome.