Why low-latency hiring is broken - and what the best firms are doing instead
Quant trading firms still hire engineers using a process designed for an era when the bottleneck was finding people. The bottleneck has changed. The process has not.
A senior C++ engineer at a top-tier prop firm in Sydney is interviewing somewhere between every quarter and every six months. Not because they are unhappy. Because the market has trained them that interviewing is the negotiation. The firm that loses them this round will lose them in two years anyway.
This is not a rant about loyalty. It is a structural observation. And it tells us that the hiring process most quant firms run - six rounds, eight weeks, three on-sites, a takehome that takes eight hours - is calibrated for the wrong constraint. It is calibrated to ration scarce candidates. The constraint has flipped. Candidates ration scarce attention.
Three things separate the firms that are winning hiring in this market from the firms that are losing.
I - Asymmetric speed in the first 72 hours
The best firms have collapsed the front of their process to a single 90-minute conversation, scheduled within 48 hours of first contact. One technical screen, one cultural conversation, one offer signal - “if the next round goes well, here is the range we are working in”. The candidate now has information that competing firms will not give them for three weeks.
This is not about being less rigorous. It is about being rigorous earlier. The full assessment still happens - but the candidate is already structurally committed before round four.
II - Pre-disclosed comp ranges
The firms that still wait until offer stage to talk comp are losing candidates they should not be losing. The strongest engineers in this market have seen enough offers to triangulate the band. They are not going to invest eight weeks finding out whether the firm is even in their range. The firms that win publish the band - privately, on the first call - and let the candidate self-select.
Some hiring managers find this distasteful. The data is unambiguous: published bands shorten time-to-close, raise offer acceptance rates, and filter out candidates who would have churned six months in.
III - Technical interviews that match the work
The firms still running LeetCode mediums in 2026 are running a process they could have lifted from a 2014 Google blog post. The engineers who do well in that process are not, on average, the engineers who do well at the work. We have placed enough of both to know.
The interviews that signal a firm worth joining now: a real production codebase, a real performance problem, a forty-minute conversation about trade-offs. Sometimes a takehome that is genuinely useful work the candidate can talk about publicly afterwards. The signal to the candidate is: this firm respects the work, and respects my time.
For clients, the implication is concrete. If your process is more than four steps and longer than three weeks, you are losing in the first ten days. The fix is not to relax the bar - it is to compress the process.
For candidates: if a firm cannot give you a clear comp band and a defined three-step process on the first call, they are not yet a serious bidder for your time. Move on. There are several that can.