STAR Method for Behavioural Interviews: 8 Full Examples (2026)

S. At Razorpay, my team of 4 backend engineers was asked to migrate our settlement service off a monolith to 14 microservices, with a hard deadline driven by our external audit (Q2 2024).

T. As the Staff Engineer on the project, I owned the migration plan, rollback strategy, and zero-downtime deployment path.

A. I broke the migration into 14 weekly phases, one service per week. I set up feature flags so each service could be toggled back to monolith traffic in 60 seconds. I ran weekly syncs with SRE, on-call rotations through each switchover, and wrote a 12-page runbook the team used for every migration.

R. We shipped all 14 services on schedule with zero revenue-impacting incidents. Deploy time dropped from 45 minutes to 4 minutes. Post-audit, the external firm flagged our migration doc as best-practice for their other clients.

S. At Swiggy, the business team wanted to launch a 10 percent blanket discount for all Instamart users during festive season. My data showed this would cannibalise our paid-subscription members.

T. As the PM for subscription, I needed to change the plan without damaging the cross-functional relationship.

A. I built a 3-slide deck showing the cohort analysis: 23 percent of discount redemptions would come from existing paying members, costing us Rs 4.2 crore in lost subscription value over 90 days. I proposed an alternative (members-only 15 percent discount, non-members 5 percent) and invited the business lead to a 30-min working session.

R. The alternative was adopted. Festive GMV grew 34 percent year-over-year, subscription renewals stayed flat (vs a projected 12 percent drop under the original plan), and the business lead and I co-presented results at the next all-hands.

S. I was lead designer on a payments redesign at a fintech startup. I pushed a visual overhaul without running it past the customer support team first.

T. The redesign shipped to 100 percent of users in one go, not staged.

A. Within 48 hours, support tickets tripled because existing users could not find the Add Funds button (I had moved it behind a two-tap flow). I rolled back in 6 hours, held a blameless post-mortem, and rebuilt the rollout plan with a 10 percent staged ramp + CS pre-brief.

R. The second release shipped 2 weeks later to 100 percent with a 4 percent drop in support tickets vs baseline. The learning: I now default to a staged rollout plus CS sync for any visual change above surface-level polish.

S. An engineer on my team was missing sprint commitments 3 sprints in a row and his PRs were being reopened by reviewers.

T. As his manager, I needed to understand the root cause and either restore performance or transition him out, without damaging team morale.

A. I had a private 45-min 1:1 to ask about blockers. Turned out he was going through a personal health issue and not flagging it. I rebalanced his load by 40 percent for 60 days, paired him with a senior mentor on architecture, and set weekly micro-goals. I also set a clear 90-day performance plan in writing.

R. He cleared the 90-day plan, shipped the auth rewrite (his largest ship to date) the following quarter, and is now one of my strongest seniors. The tradeoff: we slipped 1 mid-priority feature by 2 weeks, which I owned with my manager.

S. I joined an IB team covering Indian mid-cap IT services, but my prior experience was only in consumer tech coverage.

T. I had 3 weeks before my first model-build and pitch on a new coverage initiation (Persistent Systems).

A. I read 4 years of annual reports, listened to 12 earnings calls at 1.5x speed, set up daily 45-min calls with the sector head, and built a comparable-company model from scratch using Bloomberg + company filings. I cross-checked my assumptions with two buy-side analysts I had worked with before.

R. The Persistent Systems initiation note was published with an OUTPERFORM rating. The stock beat our target by 18 percent over the next 6 months, and 3 of our key buy-side clients increased their position within 2 weeks of publication.

S. The head of sales and I disagreed on lead-qualification criteria. He wanted every inbound demo request booked; I argued unqualified demos were wasting AE time.

T. As head of growth, I needed to find a data-backed resolution without escalating or damaging the relationship.

A. I pulled 90 days of demo-to-closed-won data. Qualified leads (using my proposed criteria) closed at 24 percent; unqualified leads at 3 percent. I built a shared dashboard so both teams saw the same numbers. I proposed a 30-day trial with stricter qualification plus a nurture track for unqualified leads.

R. Qualified pipeline grew 45 percent in the trial. CAC dropped 22 percent. Sales leader became one of my strongest advocates internally; we now co-own the qualification model.

S. We had 3 sev-1 incidents in the same week: a payment gateway outage, a data-loss bug in our mobile app, and a security CVE in a third-party library.

T. As engineering manager on-call, I had 4 senior engineers available across all 3 issues.

A. I ranked by impact and reversibility. Payment outage = customer-facing revenue loss, non-reversible = top priority (2 engineers). Data-loss bug = customer trust, partially reversible = P1 (1 engineer on rollback, 1 on fix). CVE = security exposure but no active exploit = P2 (1 engineer on patch). I set 30-min check-ins and communicated status to CEO every hour.

R. Payment restored in 47 minutes. Data-loss fix rolled out in 4 hours with 100 percent recovery of affected records. CVE patched and deployed within 24 hours (ahead of disclosure). The weekly incident post-mortem became our template for future multi-incident weeks.

S. We had to reduce our engineering team by 15 percent during a Series B bridge round. 11 engineers out of 72.

T. As Head of People, I had to design the process and communicate with each affected engineer personally.

A. I drafted a clear rubric (last-in-first-out plus skill-gap analysis), had every decision peer-reviewed by two directors, and prepared individual severance packages before any conversation. I held 1-on-1 meetings with each affected engineer within a single day, followed by company-wide announcement 2 hours after the last 1-on-1. I offered 8 weeks severance + 3 months of India health insurance + external placement support.

R. Of 11 affected, 9 had new offers within 45 days. Retention of remaining team held at 97 percent over the next 6 months (vs industry average ~70 percent post-layoff). Two of the affected engineers rehired us 18 months later when we resumed hiring.