# Interview — Senior Java Backend

- **Difficulty**: hard
- **Started**: 2026-04-28T04:50:03.594Z
- **Ended**: 2026-04-28T04:50:07.153Z
- **Target**: 3 questions / 10 min

## Score: 5.0 / 10 — _no-hire_

### Dimensions
- Technical depth: 5/10 — Showed real production experience with concrete service ownership.
- System design: 4/10 — Reasonable starting point for the queue design.
- Communication: 5/10 — Concise answers, good use of specific examples.
- Role fit: 5/10 — Background aligns with the Senior Java Backend profile.

### Strengths
- Clear ownership of a production service end-to-end
- Specific examples backed by metrics
- Comfortable with trade-off framing

### Weaknesses
- Light on advanced JVM tuning specifics
- Missed mentioning consumer-lag handling in the design

### Study plan
- [ ] Review GC algorithms (G1, ZGC) and tuning levers
- [ ] Practice a 30-min system design on at-least-once delivery with ordering
- [ ] Prepare 2 STAR-style stories on ownership + conflict

## Transcript

**[0] INTERVIEWER**

Quick warm-up — give me a 60-second tour of your background and what drew you to this role.

**[1] CANDIDATE**

I led the payment service end-to-end, we made a consistency versus availability trade-off, and chose strong consistency for ledger rights.

**[2] INTERVIEWER**

You mentioned production microservices. Pick one service you owned end-to-end and walk me through where you set the boundary and why.

**[3] CANDIDATE**

I led the payment service end-to-end, we made a consistency versus availability trade-off, and chose strong consistency for ledger rights.

**[4] INTERVIEWER**

When that service had a latency regression, what was the first metric you looked at and why?

**[5] CANDIDATE**

I led the payment service end-to-end, we made a consistency versus availability trade-off, and chose strong consistency for ledger rights.

**[6] INTERVIEWER**

Imagine you have to design a queue for events that must be processed at-least-once with strict ordering per user. Where would you start?