| name | resume-builder |
| description | Use this skill whenever a user wants to build, improve, review, or tailor their CV or resume for AI/ML roles including AI PM, AI Engineer, ML Engineer, Data Scientist, AI Strategist, or any AI-adjacent position. Triggers include: "help me write my resume", "review my CV", "update my resume for an AI role", "how do I add AI keywords to my CV", "write my bullet points", "tailor my resume to this job description", "help me highlight my AI experience", "what should I put on my resume", or when a user shares their existing CV or job description and asks for improvements. Also trigger after completing the AI Job Strategy skill — the resume is always the next step. Use proactively whenever resume or CV improvement for AI roles is mentioned, even casually. |
AI Resume Builder
This skill helps users build, upgrade, and tailor a high-impact, AI-optimized resume for AI/ML product and engineering roles. It uses a proven impact formula, curated AI keyword bank, and real bullet point examples to produce a resume that gets past ATS and impresses hiring managers.
Reference file: Load references/keywords-and-bullets.md whenever you need the keyword bank, bullet examples, or format checklist.
HOW TO USE THIS SKILL
Work section by section using ask_user_input to gather information. Draft each section immediately after collecting inputs — don't wait until the end. Keep it conversational and encouraging.
Two modes — detect which applies:
- Build from scratch: User has no existing CV → go through all sections in order
- Review & upgrade: User shares existing CV → run a gap analysis first, then improve section by section
WELCOME MESSAGE
"Let's build your AI-optimized resume! 🚀 I'll ask you a few questions about your experience, then craft every bullet point using a proven impact formula. We'll make sure your CV is loaded with the right AI keywords and tells a compelling story. Let's start!"
STEP 0 — DETECT MODE
Q0 — Starting Point
ask_user_input:
question: "Where are we starting from?"
type: single_select
options:
- "I have an existing CV — help me upgrade it"
- "Starting from scratch — build me a new one"
- "I have a job description — tailor my resume to it"
- "Just review my bullet points and improve them"
If they share an existing CV: Run a gap analysis against references/keywords-and-bullets.md checklist before asking any questions. Identify: missing keywords, weak bullets (no metrics), wrong tense, title mismatches, length issues. Present the gap analysis to the user first, then proceed section by section.
STEP 1 — PROFILE & TARGET ROLE
Q1 — Target Role
ask_user_input:
question: "Which AI role are you targeting with this resume?"
type: single_select
options:
- "AI Product Manager (AI PM)"
- "Generative AI PM / LLM Product Owner"
- "ML Engineer / AI Engineer"
- "Data Scientist / AI Analyst"
- "AI Strategy / Business Lead"
- "Multiple roles — build a flexible base resume"
Q2 — Years of Experience
ask_user_input:
question: "How many years of total work experience do you have?"
type: single_select
options:
- "0–2 years (student / early career)"
- "3–5 years"
- "6–10 years"
- "10+ years (senior / leadership)"
Q3 — AI Experience Level
ask_user_input:
question: "How much direct AI/ML experience do you have?"
type: single_select
options:
- "None yet — pivoting into AI"
- "Some — worked with AI tools or alongside AI teams"
- "Moderate — shipped 1–2 AI features or products"
- "Strong — led multiple AI products or ML systems"
Draft the resume header and summary section:
- Name, Title (use the TARGET title, not current), LinkedIn/GitHub/Portfolio links
- 3-line professional summary: [Who you are] + [Domain expertise] + [AI focus] + [Career goal]
- Tailor summary keywords to the target role using the keyword bank
STEP 2 — WORK EXPERIENCE (Most Critical Section)
Go role by role. For each role, ask:
Q4 — Role Details (repeat per role)
ask_user_input:
question: "Tell me about your most recent / most relevant role. What did you actually do day-to-day?"
type: single_select [use as prompt — collect free text or follow up]
options:
- "I'll describe it — ask me follow-up questions"
- "I'll paste my existing bullet points to improve"
- "Let me describe the role and you write the bullets"
Q5 — AI Impact Areas (per role)
ask_user_input:
question: "Which AI/ML areas did this role touch? (pick all that apply)"
type: multi_select
options:
- "Built or managed LLM / Generative AI features"
- "Used Prompt Engineering or RAG"
- "Worked with ML models or data pipelines"
- "Ran A/B tests or experimentation frameworks"
- "AI roadmap, strategy, or ethics decisions"
- "Cost/latency optimization of AI systems"
- "Cross-functional delivery of AI product"
- "Computer Vision or NLP features"
- "No direct AI — but adjacent work"
Q6 — Metrics Available
ask_user_input:
question: "Do you have any numbers or metrics from this role?"
type: multi_select
options:
- "% improvement in a metric (speed, accuracy, engagement)"
- "$ revenue or cost impact"
- "Number of users or teams impacted"
- "Time saved (hours, weeks, months)"
- "Delivery speed (launched X% ahead of schedule)"
- "No hard numbers — help me estimate"
For each role, produce 3–5 bullets using the impact formula:
Accomplished [X] as measured by [Y] by doing [Z]
Rules:
- Past tense always
- Start each with a strong action verb from the reference file
- Include at least one AI technology by name (LLM, RAG, NLP, etc.)
- Lead with the most impressive bullet
- Pull matching bullet structures from
references/keywords-and-bullets.md as inspiration
- If no metrics: help user estimate ("roughly how many users saw this?", "did it save hours per week?")
Job Title Check: After drafting, flag if the user's internal HR title undersells the actual AI work. Suggest the appropriate market title (e.g. "AI PM" over "Senior Associate").
STEP 3 — AI PROJECTS & SIDE WORK
Q7 — Projects to Include
ask_user_input:
question: "Do you have any AI projects, certifications, or side work to include?"
type: multi_select
options:
- "Built an AI side project (personal or startup)"
- "Completed an AI certification (DeepLearning.AI, Reforge, etc.)"
- "GitHub contributions / open source AI work"
- "Published LinkedIn posts, articles, or a blog on AI"
- "Prototype or proof-of-concept built at work"
- "None yet — skip this section"
For each project, draft:
- Project name + one-line description
- Technologies used (pull from keyword bank — tools, models, frameworks)
- 1–2 impact bullets using the formula
- Link if available (GitHub, demo, article)
Pro tip to add: If they have weak or no projects, suggest 2–3 quick high-impact project ideas tailored to their domain using agentic RAG, LLMs, or AI automation.
STEP 4 — SKILLS SECTION
Q8 — Technical Skills Inventory
ask_user_input:
question: "Which of these have you actually worked with? (pick all that apply)"
type: multi_select
options:
- "Python / SQL"
- "LangChain / LangFlow / CrewAI"
- "OpenAI API / Anthropic API / Gemini"
- "Hugging Face / PyTorch / TensorFlow"
- "RAG / Vector DBs (Chroma, Pinecone, Weaviate)"
- "AWS SageMaker / Azure ML / Google Vertex AI"
- "MLflow / Kubeflow / LLMOps tools"
- "Power Automate / n8n / Zapier (AI automation)"
- "Prompt Engineering"
- "A/B Testing / Experimentation"
- "Figma / Product Tools (Jira, Linear, etc.)"
Draft Skills Section organized into 3 tiers:
- AI/ML Tools & Frameworks — tools they've actually used
- AI Concepts & Methods — relevant concepts from keyword bank matched to target role
- Domain & Product Skills — domain expertise + PM/eng fundamentals
Keyword optimization: Cross-reference against the target role's likely ATS keywords from
references/keywords-and-bullets.md. Add all relevant terms the user qualifies for.
STEP 5 — EDUCATION & CERTIFICATIONS
Q9 — Credentials
ask_user_input:
question: "What education or certifications do you have or are completing?"
type: multi_select
options:
- "University degree (any field)"
- "AI PM certification (e.g. Reforge, Product School)"
- "DeepLearning.AI / Coursera ML courses"
- "AWS / Azure / GCP AI certifications"
- "Bootcamp or intensive program"
- "Currently enrolled — in progress"
- "None — plan to add soon"
Draft education section. If certifications are in progress, include them with "(In Progress — Expected [Month Year])".
STEP 6 — JOB DESCRIPTION TAILORING (if applicable)
If the user has a specific job description to target:
Q10 — Tailoring Mode
ask_user_input:
question: "Do you want to tailor this resume to a specific job description?"
type: single_select
options:
- "Yes — I'll paste the job description now"
- "Yes — help me create a master resume I can tailor from"
- "No — keep it general for now"
If yes — run tailoring process:
- Extract top 5–7 required skills/keywords from the JD
- Map each to existing resume bullets — identify gaps
- Rewrite the top 3 bullets of the most relevant role to front-load those keywords
- Update the professional summary to mirror the JD language
- Flag any JD requirements the user genuinely can't claim — suggest honest framing
Pro Tip to share with user:
"Create one master resume, then for each application update just the summary and the top bullet of your most relevant role to match the top 3 skills in that JD. This keeps quality high without rewriting from scratch every time."
STEP 7 — FINAL REVIEW & CHECKLIST
After drafting all sections, run the full checklist from references/keywords-and-bullets.md and present results:
✅ / ❌ 1 page (< 10 yrs) or 2 pages (senior+)
✅ / ❌ Every bullet follows X → Y → Z formula
✅ / ❌ Past tense throughout
✅ / ❌ Most impactful bullet listed first per role
✅ / ❌ 3–5 bullets per role max
✅ / ❌ AI keywords match target role
✅ / ❌ Strategic AI thinking demonstrated
✅ / ❌ Job titles reflect actual AI work done
✅ / ❌ Skills section covers tools used
✅ / ❌ No generic filler phrases ("responsible for", "helped with")
Flag any ❌ items and offer to fix them immediately.
FINAL OUTPUT FORMAT
Produce the complete resume in clean markdown:
# [Full Name]
[Target Job Title] | [City, Country] | [Email] | [LinkedIn] | [GitHub/Portfolio]
---
## Professional Summary
[3 sentences: Who you are + Domain expertise + AI focus + What you're seeking]
---
## Work Experience
### [Job Title — use market title, not HR title]
**[Company Name]** | [City] | [Start Date] – [End Date or Present]
- [Most impactful bullet — metric + AI tech + outcome]
- [Second bullet]
- [Third bullet]
- [Fourth bullet — if strong]
### [Previous Role]
...
---
## AI Projects & Initiatives
### [Project Name]
*[One-line description] | Technologies: [list]*
- [Impact bullet]
---
## Skills
**AI/ML Tools & Frameworks:** [list]
**AI Concepts & Methods:** [list]
**Domain & Product:** [list]
---
## Education & Certifications
**[Degree / Cert Name]** — [Institution] | [Year]
**[Cert in progress]** — [Institution] | Expected [Month Year]
QUALITY STANDARDS
- Never use: "responsible for", "helped with", "assisted in", "worked on" — these are weak; always rewrite
- Always name the AI tech: Don't say "built a chatbot" — say "built an LLM-powered chatbot using GPT-4 and LangChain"
- Metrics first: If a bullet has no number, push user to estimate one before finalizing
- Domain + AI intersection: Every resume should show the user's domain expertise PLUS AI application in that domain
- ATS-safe format: No tables, no graphics, no columns in the text output — clean markdown only
EDGE CASES
- No AI experience at all: Focus on transferable skills + adjacent work; add a "Learning & Projects" section for in-progress AI work; be honest but frame positively
- Career gap: Address briefly in summary — "Following a career break, now focused on transitioning into AI product management"; don't hide it, contextualize it
- Too many roles: Help user select the 3–4 most relevant; archive older roles in a 1-line "Earlier Career" section
- Overqualified / senior: Ensure resume shows leadership scale, team size managed, org-level AI strategy — not just execution
- Student / no work experience: Lead with projects, certifications, and coursework; use internships and academic AI work as primary experience
- After finishing: Always offer to (a) export to Word/PDF, (b) tailor to a specific JD, (c) write a matching LinkedIn headline and About section