Resumes
Software Engineer Resume Bullets That Get Interviews
A simple formula for engineering resume bullets, before-and-after rewrites for backend, frontend, DevOps and data roles, and how to find numbers when you think you have none.
Updated · 8 min read
Most engineering resumes describe responsibilities: "Worked on the payments service." Recruiters and hiring managers are looking for outcomes: what changed because you were there. A single structure fixes most weak bullets.
The formula
Action verb + what you built or changed + key technologies + measurable result. Not every bullet needs all four, but your top three bullets in each role should. Put the result early when it is strong: "Cut p95 latency from 900 ms to 180 ms by…" reads faster than burying it at the end.
Before and after
Backend
- Before: Responsible for the order API.
- After: Redesigned the order API's Postgres queries and added Redis caching, cutting p95 latency from 900 ms to 180 ms at 3k requests/second.
Frontend
- Before: Built React components for the dashboard.
- After: Rebuilt the analytics dashboard in React and TypeScript with code-splitting, reducing initial load from 4.1 s to 1.6 s and support tickets about slowness by half.
DevOps / SRE
- Before: Managed CI/CD pipelines and AWS infrastructure.
- After: Moved 30 services from hand-built EC2 to Terraform-managed EKS with GitHub Actions, taking deploys from weekly to 20+ per day with zero-downtime rollbacks.
Data engineering
- Before: Maintained ETL jobs.
- After: Replaced 40 nightly batch jobs with incremental dbt models on Snowflake, cutting warehouse spend 28% and making dashboards refresh hourly instead of daily.
Machine learning
- Before: Worked on recommendation models.
- After: Shipped a PyTorch two-tower retrieval model behind an A/B test that raised click-through 6.2%, and built the offline evaluation pipeline the team now uses for every model.
Finding numbers when you think you have none
- Time: build time, deploy frequency, page load, query latency, time to onboard a new engineer.
- Scale: requests per second, rows per day, users, services, regions.
- Money: cloud cost saved, revenue on a feature, licences retired.
- Quality: incidents, error rates, test coverage, tickets closed.
- People: engineers mentored, teams that adopted your tool, interviews conducted.
Estimates are fine when they are honest and defensible ("about 30%"). Interviewers ask how you measured it, so know the story behind every number.
Common mistakes
- Starting every bullet with "Responsible for" or "Helped".
- Listing tools without saying what you did with them.
- Six-line bullets: keep each to one or two lines.
- Old roles with as much detail as your current one — trim anything older than about ten years.