AI resume builder and marketplace onboarding for a hiring platform
Applying for jobs on a hiring platform meant uploading a finished resume, which held back candidates who had the skills but not the document. The team needed AI-generated resumes from plain descriptions, a live editor, and onboarding for a new agent marketplace, all in the same delivery window.
- My role
- MERN Stack Developer - feature development and code review
- Team
- Full production team - developers, designers, product managers
- Timeline
- 3 months to ship new features, with ongoing bug fixes
- Constraint
- AI-generated resumes had to be consistent, structured and immediately editable - unreliable output would break the application flow
The problem
Requiring a polished, ready resume was a barrier for candidates who had the skills but not the document. The goal was for a candidate to describe what they do and get a professionally structured resume they could edit, download or submit straight away.
At the same time, the company was launching an agent marketplace and needed onboarding for buyers and sellers, with separate logic and screens, built alongside the resume work.
The tradeoff I made
Instead of generating a resume and handing back a static document, I built the output directly into a live editor. The candidate sees the resume take shape and can refine it immediately, with no download, re-upload or switching tools. The AI layer and the editor had to be designed together from the start, but the result is one smooth journey: describe yourself, watch the resume build, edit it and apply.
A complete user journey over a technically simpler feature drop.
Outcomes
Full resume generated from a plain text description, replacing a multi-step manual upload
Onboarding flows shipped for buyers and sellers - sign up, login, store or profile setup, and routing to their dashboards
Both features delivered on schedule alongside ongoing bug fixes
What I would do differently
I would agree a file size limit from the start. The resume editor grew quickly and too much logic ended up in the same files, so refactoring later took time away from features. An early limit would have forced cleaner separation before it became painful.