CV Builder

Overview

Generate professional, well-formatted CVs and resumes from structured YAML input using RenderCV (v2, Typst-based). Define your experience, education, skills, and projects in a YAML file and produce publication-ready PDF output with consistent formatting. Choose from nine built-in themes or create a custom one with rendercv create-theme.

Instructions

When a user asks you to create or update a CV/resume, follow these steps:

Step 1: Gather career information

Ask the user for or extract the following details:

• Personal info: Full name, email, phone, location, LinkedIn/GitHub URLs
• Summary: A brief professional summary or objective (2-3 sentences)
• Experience: Job titles, companies, dates, bullet-point accomplishments
• Education: Degrees, institutions, graduation dates, GPA (optional)
• Skills: Technical skills, languages, tools, frameworks
• Projects (optional): Notable projects with descriptions
• Certifications (optional): Professional certifications with dates

Step 2: Create the YAML input file

Install RenderCV (v2.x, needs Python 3.12+) and let it scaffold a valid starting file:

pip install "rendercv[full]"
rendercv new "Jane Smith"        # writes Jane_Smith_CV.yaml

Then fill it with the user's data. The schema is strict: unknown keys are rejected, so keep to these names:

cv:
  name: Jane Smith
  headline: Senior Software Engineer
  location: San Francisco, CA
  email: jane.smith@gmail.com
  phone: "+14155550123"          # must be a valid phone number, E.164 style is safest
  website: https://janesmith.dev
  social_networks:
    - network: LinkedIn
      username: janesmith
    - network: GitHub
      username: janesmith
  sections:                       # section titles are free text; order is kept
    summary:
      - Senior software engineer with 8 years of experience building scalable web applications in Python, TypeScript, and cloud infrastructure.
    experience:
      - company: Northwind Systems
        position: Senior Software Engineer
        location: San Francisco, CA
        start_date: 2021-03
        end_date: present
        highlights:
          - Led migration of a monolith to microservices, cutting deploy time by 70%
          - Designed a real-time pipeline processing 2M events/day
    education:
      - institution: University of California, Berkeley
        area: Computer Science
        degree: BS
        start_date: 2014-09
        end_date: 2018-05
        highlights:
          - "GPA: 3.8/4.0"
    skills:
      - label: Languages
        details: Python, TypeScript, Go, SQL
      - label: Infrastructure
        details: AWS, Docker, Kubernetes, Terraform

design:
  theme: sb2nov
  page:
    size: us-letter               # or a4
    top_margin: 2cm
    bottom_margin: 2cm
    left_margin: 2cm
    right_margin: 2cm
  typography:
    font_size:
      body: 10pt

A plain string in a section is a text paragraph; a mapping with company/position is an experience entry, institution/area an education entry, and label/details a one-line skills entry. The older v1 keys (design.font_size, design.page_size, design.margins) no longer exist.

Step 3: Choose a theme

Built-in themes in v2.8: classic, sb2nov, moderncv, engineeringresumes, engineeringclassic, harvard, ink, opal, ember. Set one in design.theme. classic and engineeringresumes are safe choices for conservative or ATS-parsed applications; render the same file with two or three themes and compare the PNGs when the user is undecided.

Step 4: Generate the PDF

rendercv render Jane_Smith_CV.yaml

RenderCV compiles with Typst (no LaTeX needed) and writes rendercv_output/:

• Jane_Smith_CV.pdf - the final resume
• Jane_Smith_CV.typ - Typst source for manual tweaks
• Jane_Smith_CV_1.png - one PNG per page, useful for a quick visual check
• Jane_Smith_CV.md and Jane_Smith_CV.html - text versions

Useful flags: -o <folder> changes the output folder, -nopng / -nomd / -nohtml skip files, -w re-renders whenever the YAML changes, and --cv.phone "+14155550123" overrides a single field without editing the file.

Step 5: Validate and iterate

Validation runs on every render and prints a table of field, value and explanation for each problem, so fix the listed fields and re-run. Then review the PDF with the user:

• Check for text overflow, wrapped dates and a spilled second page
• Verify dates and details are correct
• Adjust theme, design.page margins or design.typography.font_size if needed

Examples

Example 1: Create a new resume from scratch

User request: "Help me create a professional resume. I'm a data scientist with 5 years of experience."

Steps:

1. Ask the user for their full career details
2. Run rendercv new "Priya Raman", then fill Priya_Raman_CV.yaml using the sb2nov theme
3. Run rendercv render Priya_Raman_CV.yaml
4. Present the PDF output path and offer to adjust formatting

Output:

Created Priya_Raman_CV.yaml with your career data
Generated PDF: rendercv_output/Priya_Raman_CV.pdf
Also available: .typ, .md, .html and a PNG preview per page

The resume uses the sb2nov theme.
Would you like to adjust the template, font size, or content?

Example 2: Update an existing CV with a new job

User request: "Add my new position at Google to my existing CV"

Steps:

1. Read the existing Priya_Raman_CV.yaml
2. Close the previous job (end_date: 2026-09) and add the new entry at the top of experience with start_date: 2026-10, end_date: present
3. Re-run rendercv render Priya_Raman_CV.yaml
4. Confirm the update with the user

Output: Generated PDF: rendercv_output/Priya_Raman_CV.pdf, still two pages.

Example 3: Switch template theme

User request: "Change my resume to a more traditional format for a finance role"

Steps:

1. Read the existing Priya_Raman_CV.yaml
2. Change design.theme from sb2nov to classic
3. Optionally set design.typography.font_size.body: 11pt for readability
4. Re-run rendercv render Priya_Raman_CV.yaml

Output: a regenerated PDF in the classic layout; the content is untouched.

Guidelines

• Always use quantifiable achievements in bullet points (numbers, percentages, metrics).
• Keep the resume to 1 page for candidates with under 10 years of experience; 2 pages maximum for senior professionals.
• Use action verbs to start each bullet point (Led, Built, Designed, Implemented, Reduced).
• Order experience entries reverse-chronologically (most recent first).
• Tailor the skills section to match the target role when the user specifies one.
• RenderCV validates the whole file on render; quote values that YAML could misread ("GPA: 3.8/4.0", phone numbers).
• If rendercv is not installed, guide the user: pip install "rendercv[full]" (Python 3.12+; the full extra bundles Typst and fonts).
• Files written for RenderCV v1 (design.font_size, design.page_size, .tex output) fail validation on v2: migrate them rather than pinning an old version.
• Store the YAML file in version control so the user can track resume changes over time.
• Avoid including personal photos, age, or marital status unless the user explicitly requests it and it is customary for their region.