AI Resume Generator — Automated CV Pipeline via n8n

Challenge

A career coaching company needed to scale its resume writing service without scaling its staff headcount. Each CV required a human writer to interview the client, structure their experience, tailor the language per industry, and format the document. Turnaround was 2–3 days; throughput was capped by writer capacity. The goal: automate 80% of the production pipeline while keeping output quality high enough that clients couldn't tell the difference.

Options Considered

1. Off-the-shelf resume builder SaaS (Canva, Zety) — templates only, no AI rewriting of raw client input, no API for automation integration. Rejected.
2. Custom-coded pipeline (Node.js or Python) — full control, but significant dev overhead to build the trigger, orchestration, retry logic, and monitoring from scratch. Slower to iterate on workflow changes.
3. n8n workflow + GPT-4 + PDF generation — chosen. n8n handles triggers, routing, retries, and integrations visually; GPT-4 rewrites raw input into polished, role-targeted copy; a PDF node assembles the final document. Low maintenance, fast to modify.

Decision

The client submits an intake form (work history, skills, target role, tone preference). The form submission triggers an n8n workflow. GPT-4 receives a structured prompt that instructs it to rewrite the raw input as professional resume copy, tailored to the target role, with ATS-friendly keyword density. The rewritten content is passed to a templating node that renders it into an HTML resume template. Puppeteer converts the HTML to PDF. The final document is uploaded to cloud storage and a download link is emailed to the client automatically.

Implementation

n8n self-hosted on a VPS handles the full orchestration: Webhook node receives the form payload, HTTP Request nodes call the OpenAI API, a Code node post-processes the JSON response into template variables, an HTML node renders the resume layout, a Puppeteer node (via custom n8n node) renders to PDF, and an S3 node uploads the file. A final Send Email node delivers the link. Error branches route failed jobs to a Slack alert for human review. The entire workflow is version-controlled as a JSON export.

Outcome

Turnaround dropped from 2–3 days to under 10 minutes. Writer time per resume dropped from ~2 hours to a 5-minute review pass on the generated draft. The pipeline handles dozens of submissions per day without additional staff.