JOBSPRINT OS
Turn a scattered job search into a deliberate workflow.
A personal product for bringing opportunities, resumes, project evidence and interview preparation into one workspace. Built iteratively with AI assistance, starting from a workflow I needed myself.
The tabs were not the system.
A job search quickly becomes a collection of documents, email threads and repeated decisions. The missing piece was connected context: which opportunity mattered, which experience supported it, which resume was selected and what needed to happen next.
JobSprint OS brings that context into a shared workspace instead of treating every action as an isolated task.
The decisions I drove.
- A shared opportunity pipeline with explicit stages and next actions.
- A project library and resume selection workflow, keeping supporting experience close to the opportunity.
- Interview preparation that brings company research and relevant topics together.
- Approval boundaries and activity evidence for actions outside the product.
- Portable hosting and persistent storage, separate from an individual chat session.
A product beyond a conversation.
The application uses React and TypeScript, with Cloudflare Workers for hosting and application services. Cloudflare D1 stores operational records, R2 stores files, and Supabase Auth handles authentication.
Keep preparation separate from execution.
A drafted message is not a sent message. A prepared application is not a submitted application. The workflow distinguishes these states so that AI assistance supports decisions without quietly turning intentions into claimed results.
Product requirements, iterations and review were driven by my own use. Implementation was carried out with AI assistance, with particular attention to the approval workflow and continuity of records.
Useful without becoming a public data leak.
The operational workspace contains private job-search records, so it is not exposed as an open demo. This public case study describes the product and architecture, not personal applications, recruiter messages or other users' information.
The work demonstrates product thinking and iterative delivery. I do not attribute hiring outcomes to the software without evidence.
Let's talk about the product ↗