STAR Story Bank — Google 5 YOE#

Use STAR+R: Situation → Task → Action → Result → Reflection.

Write these from your real backend experience (PostgreSQL, Bigtable, BigQuery, Django/FastAPI, Spring Boot).

---

1. Ambiguity#

Situation: (project, unclear scope)
Task: (your specific job)
Action: (how you reduced ambiguity — asked, proposed, decided, set)
Result: (metric)
Reflection: (what you'd repeat)

Example skeleton — fill with your real project:

On a BigQuery analytics migration, the data model was undefined and stakeholders had conflicting SLAs. I drove a 1-week alignment: defined event schemas, partitioning strategy, and query budgets. Shipped 3 weeks early with 40% cost reduction. I'd repeat the schema-first approach.

---

2. Disagreement / Pushback#

Situation: (technical disagreement with PM, peer, or leadership)
Task: (your position)
Action: (how you pushed back with data, or accepted and delivered)
Result: (what changed)
Reflection:

A PM wanted to ship a BigQuery query to the serving path. I showed query-time and concurrency numbers; we moved to an async materialized view. Query went from 12s to 200ms. I'd do the same pushback again.

---

3. Cross-functional collaboration#

Situation: (PM, designer, ops, data eng, etc.)
Task: (your role bridging teams)
Action: (how you aligned)
Result: (delivered)
Reflection:

---

4. Failure / production incident#

Situation: (what broke)
Task: (your responsibility)
Action: (what you did, owned, fixed)
Result: (time to restore, what changed)
Reflection:

Deadlocks in PostgreSQL under load during a schema migration. I owned the postmortem, added retry + exponential backoff, and introduced a migration gate. Post-incident query latency dropped 60%. I'd have flagged the lock timeout earlier in design.

---

5. Mentorship / raising the bar#

Situation: (junior joining, code quality, review culture)
Task: (your role)
Action: (what you did)
Result: (impact)
Reflection:

---

6. Impact / ownership (project shipped)#

Situation: (what you built)
Task: (your scope)
Action: (how you delivered)
Result: (metrics — latency, cost, throughput, users)
Reflection:

---

7. Impact 2 (second project, different domain)#

Same skeleton. Different domain (analytics, infra, migration, performance).

---

8. Technical depth (hard problem)#

Situation: (tough technical problem — query optimization, distributed system, data model)
Task: (solve it)
Action: (deep technical reasoning)
Result: (numbers)
Reflection:

Optimized a BigQuery ETL from 45 min to 8 min: partitioned by ingestion date, clustered on user_id, replaced nested loops with window functions. Saved $24k/month in compute.

---

STAR rules#

1. First person, present tense ("I led", not "we").
2. Metric in every story (time, cost, %, users).
3. Never blame a company, team, or person.
4. Keep to 3–4 min per story when told aloud.
5. Prepare 2 versions: one 60s (resume) + one 3min (full).
6. Rehearse aloud 8 min each. Record yourself.
7. Never say "I don't have a story for that." Pick the closest theme.

---

Theme quick map#

---

Prep this week (do now)#

1. Fill in all 8 tables with real projects from your backend experience.
2. Pick your 5 strongest stories — those are your interview anchors.
3. Time each aloud. Trim to ≤ 3 min.
4. Record one full run-through. Listen for "um" and vagueness.