Engineering Hiring Framework — 200 Interviews to 40 Quality Hires

The Problem

At a fast-scaling fintech, the engineering organisation needed to grow from 80 to 120 engineers in a single quarter to support three new product lines. The hiring process was not ready for that kind of volume.

Every interviewer had their own approach. Some asked LeetCode-style algorithm puzzles. Others had free-form technical conversations. A few just asked "tell me about yourself" for 45 minutes. There were no rubrics, no calibration sessions, and no consistent scoring. Two interviewers could evaluate the same candidate and reach completely opposite conclusions with no way to reconcile.

The result was slow, inconsistent, and expensive. Time-to-hire averaged 45 days. The offer-decline rate was 30% — nearly one in three candidates who received offers chose to go elsewhere, often citing a disorganised interview experience. Hiring managers were frustrated, recruiters were overwhelmed, and the talent pipeline leaked at every stage.

What I Did

I treated the hiring process as a product and redesigned it end to end.

First, I created standardised interview rubrics for each role level — junior, mid, senior, and staff engineer. Each rubric defined the specific competencies being evaluated (technical depth, system design, collaboration, problem-solving) with clear scoring criteria on a 1-4 scale. Interviewers assessed specific competencies, not general vibes.

Second, I redesigned the technical assessment. We replaced the generic algorithm test with a take-home exercise that reflected actual work the candidate would do on the job. For backend roles, it was a small API design and implementation. For frontend roles, a component-building exercise. The exercises were scoped to take 2-3 hours and came with clear evaluation criteria.

Third, I trained every interview panellist. We ran two calibration sessions where interviewers scored the same mock candidate and discussed differences. This aligned the team on what "strong" versus "adequate" actually looked like. I also introduced structured debrief sessions — instead of going around the room sharing opinions, we used a written scorecard submitted before the debrief to prevent anchoring bias.

Fourth, I focused on candidate experience. We reduced the interview process from five rounds to three. Every candidate received a prep guide explaining what to expect. Scheduling was consolidated so candidates could complete all interviews within one week. Recruiters sent status updates at defined intervals so candidates never felt forgotten.

Finally, I established a hiring dashboard tracking funnel metrics — applications, screens, interviews, offers, and acceptances — with conversion rates at each stage. This gave us real-time visibility into where the funnel was leaking.

The Outcome

Time-to-hire dropped from 45 days to 22 days. The offer acceptance rate improved from 70% to 85%, driven primarily by the improved candidate experience and faster process. We made 40 hires in the quarter — hitting the target exactly.

Quality metrics held strong: 90-day retention for the new cohort was 95%, and hiring manager satisfaction with new hires scored 4.3 out of 5. The structured rubrics also reduced bias concerns — we saw a measurable improvement in diversity of hires compared to the previous two quarters. The framework became the standard across all engineering hiring and was later adapted for product and design roles.