Skip to content

Latest commit

 

History

History
307 lines (238 loc) · 13.5 KB

File metadata and controls

307 lines (238 loc) · 13.5 KB

CV & Cover Letter Evaluation Rubric — All Tech Roles

Use this rubric for any full evaluation of a CV or cover letter. The scoring framework is role-adaptive: core dimensions stay the same, but weights shift based on role family. Follow it as a checklist; produce a score per dimension and a global score.


Role Family Detection

Before scoring, detect the role family from the CV content or target JD:

Role Family Detect When
Software Engineer Frontend, Backend, Full-stack, Web, Mobile, Embedded
Data & ML Data Engineer, Data Scientist, ML Engineer, AI Engineer
DevOps / SRE / Platform DevOps, SRE, Platform Engineer, Cloud Engineer, Infrastructure
Design UX, UI, Product Designer, Interaction Designer, Design Systems
Security Security Engineer, AppSec, Pentester, SOC, CISO
Engineering Management Engineering Manager, Tech Lead, VP Engineering, CTO
TPM Technical Program Manager, Program Manager (technical)
QA / SDET QA Engineer, SDET, Test Automation, Quality
Solutions / DevRel Solutions Engineer, Sales Engineer, Developer Advocate, DevRel
Product Manager PM, Product Manager, AI PM, Group PM, Director of Product

Dimension Weights by Role Family

Dimension SWE Data/ML DevOps/SRE Design Security EM/Lead TPM QA/SDET SE/DevRel PM
Shipped Evidence 30% 25% 25% 20% 25% 25% 20% 25% 20% 30%
Quantified Impact 25% 25% 25% 15% 20% 25% 25% 20% 20% 20%
Tech/Tool Visibility 15% 15% 15% 10% 15% 10% 10% 15% 15% 15%
ATS Compatibility 10% 10% 10% 5% 10% 10% 15% 10% 10% 15%
Keyword Match 10% 10% 10% 10% 15% 10% 15% 15% 10% 10%
Public Proof Surface 10% 15% 15% 40% 15% 20% 15% 15% 25% 10%

Why Design has 40% Public Proof: For designers, the portfolio IS the application. A designer with a brilliant portfolio and mediocre CV will get interviews. A designer with a great CV and no portfolio will not.

Why SE/DevRel has 25% Public Proof: Communication IS the job. Blog posts, talks, open-source contributions are primary screening signals.


CV Rubric — Full Detail

1. Shipped Evidence

What it measures: Real, in-production work with named users/customers, real outcomes, and specific technologies.

Score What it looks like
5 Multiple production-shipped projects with named users/customers, real outcomes, and technologies explicitly named. Evidence of end-to-end ownership.
4 At least one major shipped project with quantified scope; one or more additional shipped projects.
3 Some shipped work, but unclear ownership. Hard to tell who did what.
2 Mostly tutorials, courses, certifications. Demo projects but no real users/traffic.
1 No evidence of shipping; only theory, certifications, or aspirational language.

Role-specific signals:

Role What "shipped" means
SWE Code in production, serving real users/traffic, with uptime and scale numbers
Data/ML Models in production, pipelines running reliably, dashboards used for decisions
DevOps/SRE Infrastructure running at scale, incident response stories, platform adopted by teams
Design Shipped features with before/after metrics, not just mockups
Security Vulnerabilities found, incidents handled, compliance achieved
EM/Lead Teams scaled, projects delivered, organizational outcomes
TPM Programs completed, migrations executed, cross-team deliveries
QA/SDET Test frameworks built and adopted, quality improvements measured
SE/DevRel POCs delivered, developer adoption driven, community built
PM Features shipped, metrics moved, products launched

Red flags (all roles): "Familiar with..." / "Knowledge of..." with no demonstrated build.

2. Quantified Impact

What it measures: Numbers in every bullet — scope, speed, adoption, savings, volume, reliability.

Score What it looks like
5 Every bullet has at least one number. Mix of scope, speed, adoption, savings, or volume.
4 70%+ of bullets quantified; the rest are clearly descriptive context.
3 Half the bullets have numbers; the rest are vague.
2 Few numbers; mostly "responsible for," "worked on."
1 Almost no numbers anywhere.

Quantification categories by role:

Category SWE DevOps/SRE Data/ML Design EM QA
Scale Users served, requests/day, QPS Services managed, deploys/week Rows processed, models served Users affected, sessions Team size, headcount Test cases, coverage %
Speed Latency reduction, build time MTTR, deploy time Pipeline runtime, query speed Time-to-task, flow time Cycle time, velocity Test runtime, feedback loop
Reliability Uptime %, error rate SLO attainment, incident count Pipeline uptime, data quality Error rate, accessibility On-time delivery Defect escape rate
Cost Infra cost saved Cloud spend reduced Compute optimized Design iteration saved Budget managed Manual testing reduced
Adoption Feature usage, DAU Platform users, dev adoption Dashboard users, model consumers Design system adoption Retention, promotions Framework adoption

3. Tech/Tool Visibility

What it measures: Specific tools, frameworks, and technologies named — not generic categories.

Score What it looks like
5 Has a dedicated skills/tools section AND specific names appear inside experience bullets ("Built X using React + PostgreSQL + Redis").
4 Tools named in either skills or experience but not both.
3 Tools mentioned generically ("cloud platforms," "modern frameworks").
2 "Familiar with various tools" or vague mention.
1 No specific tools visible.

Role-specific tool expectations:

Role Must-name tools (2026 baseline)
Frontend SWE React/Vue/Angular, TypeScript, Next.js/Nuxt, CSS framework, build tools
Backend SWE Language (Go/Python/Java/Node), framework, database, message queue, API style
Full-stack Both frontend + backend stacks, deployment target
Data Engineer Spark/dbt, cloud DW (Snowflake/BigQuery/Databricks), orchestrator (Airflow), streaming
Data Scientist Python, SQL, statistical framework, visualization tool, experimentation platform
ML Engineer PyTorch/TensorFlow, serving framework, MLOps tools, cloud ML platform
AI Engineer LLM APIs, RAG framework, vector DB, evaluation tools, agent framework
DevOps/SRE Container orchestration, IaC tool, CI/CD, observability stack, cloud provider
Platform Eng IDP tools, golden paths, service catalog, developer experience tooling
Designer Figma (baseline), prototyping tool, research tool, analytics integration
Security SAST/DAST tools, SIEM, cloud security tools, compliance frameworks
EM/Lead Not expected to list tools; technical credibility shown through architecture language
TPM Project tools (JIRA/Linear), documentation tools, dashboards
QA/SDET Test framework (Playwright/Cypress), CI integration, API test tool
SE/DevRel Demo environments, documentation tools, the product they sell/advocate
PM Analytics (Amplitude/Mixpanel), prototyping (Figma), project (Linear/JIRA)

AI tooling bonus (all roles): +0.5 points if the candidate demonstrates AI tool usage in their workflow:

  • Engineers: Cursor, Claude Code, GitHub Copilot, Continue, Aider
  • Designers: Midjourney, Galileo AI, AI prototyping tools
  • Managers: AI for planning, writing, analysis
  • QA: AI-assisted test generation
  • DevOps: AI for incident triage, runbook generation

4. ATS Compatibility

Run this check explicitly:

Check Pass condition
Single column? Yes
Tables in experience section? No
Standard headings? (Experience, Skills, Education, Projects) Yes
Date format consistent? "Jan 2022 -- Present" (or similar, applied uniformly)
Special characters? No smart quotes, no em-dashes in critical fields, ASCII safe
Image-based PDF? No (text must be selectable)
Header/footer with key info? Headers/footers risk being skipped — keep contact info in body
Contact info parseable? Email, phone, LinkedIn URL on separate lines, plain text

Score = number of passes / total checks x 5.

Designer exception: Designers often use visually rich CVs. Score ATS on the "plain text" version they should have alongside their portfolio. If applying through ATS, the plain version matters. If portfolio link is the primary path, ATS matters less.

5. Keyword Match

When a JD is provided:

  1. Extract 15-20 keywords from the JD (skills, tools, frameworks, soft skills)
  2. Mark which appear in the CV (count exact matches)
  3. Score = (matches / target) x 5, where target = 12 for senior, 10 for mid, 8 for junior

When no JD is provided, score against the role family's keyword set from RESEARCH.md.

Important: Match terminology exactly. If the JD says "Kubernetes," the CV should say "Kubernetes" not "K8s." Include both when possible: "Amazon Web Services (AWS)."

6. Public Proof Surface

Score What it looks like
5 Active across 3+ relevant channels with quality content matching the role.
4 Two channels active with quality content.
3 One channel polished (usually LinkedIn) but no depth elsewhere.
2 Stale or minimal public footprint.
1 No public proof.

What counts per role family:

Role Primary proof channel Secondary channels
SWE GitHub (pinned repos with READMEs) Blog, Stack Overflow, X
Data/ML GitHub + notebook demos Kaggle, blog, Hugging Face
DevOps/SRE GitHub (IaC repos, tools) Blog, conference talks
Design Portfolio website (MANDATORY) Dribbble, Behance, Medium
Security GitHub, CTF rankings, bug bounty Blog, conference talks
EM/Lead LinkedIn (detailed, active) Blog, conference talks, newsletter
TPM LinkedIn (project descriptions) Blog, internal case studies
QA/SDET GitHub (test frameworks) Blog, QA community contributions
SE/DevRel GitHub + blog + talks YouTube, podcast, newsletter
PM LinkedIn + shipped product demos Blog, side projects, X

Cover Letter Rubric

1. Hook Quality (25%)

Score What it looks like
5 Opens with something specific the company recently did, said, or shipped. Could only have been written for THIS company.
4 Opens with a specific point about the role or team.
3 Opens with candidate background tied to the role.
2 Generic but not painfully cliche.
1 "I am writing to express my interest..." or similar.

2. Connection (20%)

Score What it looks like
5 Bridges the company's specific need to one named, quantified proof point from candidate's experience.
4 Connects company need to candidate experience but proof is vague.
3 Mentions both but doesn't bridge them.
2 Restates the JD without connecting to candidate.
1 No connection visible.

3. Proof (25%)

Score What it looks like
5 2-3 quantified proof points with specific tools and outcomes.
4 One strong proof + one weaker.
3 One proof point only.
2 Vague claims of impact.
1 No quantified proof.

4. Voice (15%)

Score What it looks like
5 Sounds like a real person. First-person specific, occasional warmth, no buzzword soup.
4 Mostly human with one or two corporate phrases.
3 Acceptable but bland.
2 Reads like AI-generated boilerplate.
1 Pure AI-generated cliche.

5. Ask (10%)

Score What it looks like
5 Specific ask: "I'd value 30 minutes to discuss [specific challenge]." Includes follow-up path.
4 Clear ask but generic ("I'd love to discuss this opportunity").
3 Ask present but weak.
2 Vague closing.
1 No ask, just "Looking forward to hearing from you."

6. Length (5%)

Score What it looks like
5 250-350 words, 3 focused paragraphs
4 350-450 words
3 200-250 or 450-500 words
2 <200 or 500-600 words
1 Way too short or too long

Global Score Interpretation

Score Recommendation
4.5-5.0 Strong. Ready to apply.
4.0-4.4 Good. Minor polish recommended before sending.
3.5-3.9 Targeted rewrites needed in 2-3 sections.
3.0-3.4 Major rewrite required. List the top 3 issues.
Below 3.0 Recommend NOT applying yet. Rebuild before submitting.

Output Template

When you complete an evaluation, format it as:

## CV Evaluation: [Candidate / Role Family]

**Global Score: X.X/5**
**Role Family:** [detected]
**Weight Profile:** [role-specific weights applied]

| Dimension | Weight | Score | Notes |
|---|---|---|---|
| Shipped Evidence | X% | X/5 | ... |
| Quantified Impact | X% | X/5 | ... |
| Tech/Tool Visibility | X% | X/5 | ... |
| ATS Compatibility | X% | X/5 | ... |
| Keyword Match | X% | X/5 | ... |
| Public Proof Surface | X% | X/5 | ... |

### What Works (Top 3)
1. ...
2. ...
3. ...

### What Needs Fixing (Ranked by Impact)
1. **[Issue]** -- Why it matters -> How to fix
2. ...
3. ...

### Top 3 Rewrites (Before -> After)
1. **Before:** [exact text from CV]
   **After:** [rewritten version with quantification + tools]
2. ...
3. ...

### Recommendation
[Apply / Polish first / Major rewrite / Don't apply yet]