Zelume
Log inCreate resume
BuilderTemplatesHow it worksFree ToolsPricingBlogLog inCreate resume

Illustrative example · fictional profile, real scoring rubric

Machine learning engineer example

82
out of 100
Resume Readiness Score
Medium confidence · 3 repos · 3 READMEs

Ken, here's what we found

Ken, the profile is unusually production-shaped for an ML engineer: a serving layer with batching, a RAG pipeline that ships with its own eval harness, and a scheduled embedding job.

The eval harness on rag-lab is the standout, because it shows you measure quality instead of eyeballing it.

100
Project Quality
50
Technical Breadth
100
Documentation
67
Engineering Practices
90
Career Presentation
Turn this into a complete resume →

This is a fictional example. Run your own username above.

Standout project

serveml

The one to lead your resume with. It checks every box a recruiter skims for.

Substantial codebaseWell documentedDeployment-ready (Docker or infrastructure code)Engineering practices visible (tests or CI)External adoption or collaboratorsActively maintained

Top repositories

rag-lab★★★★substantial · documented
embed-batch★★★documented · CI

Read like a code review

+Your deployment experience is visible: containerized or infrastructure-defined projects read as production-minded.
+Most of your top projects are documented, which is rarer than it should be.
+Recent, maintained work: your most important repositories show current activity.
+Automated pipelines show up in your work, a strong engineering-practice signal.
→Presentation looks solid; the next lever is measurable impact statements in each README.

Strongest languages

Python
98%
Dockerfile
2%

Detected stack

Frameworks
FastAPI
Data & infrastructure
Docker
CI/CD
GitHub Actions

Evidence used

Everything above is computed from this. Nothing else.

3repos analyzed
3READMEs
1Dockerfiles
3CI workflows
97stars
154 KBof code
What a recruiter probably sees
"An ML engineer who owns the serving and evaluation side, not just training, and who treats latency and correctness as first-class."

Based on the public repositories analyzed, a recruiter is likely to infer an ML engineer who owns the serving and evaluation side, not just training, and who treats latency and correctness as first-class. Publishing a benchmark number or two from the eval harness would make the rigor visible at a glance.

@tensor-tinkerer · analyzed Jul 20, 2026zelume.io/github-resume

Recruiter visibility

83%

of expected backend signals present · heuristic

Missing: A database

Drafted from the evidence

Written only from the repository evidence. No invented numbers.

servemlBuilt a model serving layer with request batching on FastAPI and Triton, containerized with tests and CI
rag-labImplemented a retrieval-augmented generation pipeline with an automated evaluation harness

Where it lands in the editor

Ken Watanabe
github.com/tensor-tinkerer
Projects
  • Built a model serving layer with request batching on FastAPI and Triton, containerized with tests and CI
  • Implemented a retrieval-augmented generation pipeline with an automated evaluation harness
Skills
Python, Dockerfile, FastAPI, Docker, GitHub Actions
Experience
  • Software Engineering Intern, Freshbooks (Summer 2025)

That's the free analysis. The resume takes one more click.

  • Complete resume from all projects
  • ATS optimization
  • Tailoring to a posting
  • Cover letter
  • Full editor
Turn this into a complete resume →

Run your own username →