RoleMark

The first version’s scoring and tailoring dashboard.

Note: RoleMark is being re-architected and rebuilt from scratch. The first version shipped, is no longer running, and its codebase is archived. The current repository is a ground-up rewrite on a new stack, carrying forward the product thesis below.

Project Overview: RoleMark is built around a simple idea: every role deserves the right resume. Job seekers upload a resume and job description to get an instant compatibility score with actionable feedback, then generate role-specific resume versions tuned for ATS systems and hiring managers. The product also includes AI-assisted cover letters, a rich editing experience, and workflow tools to stay organized across applications.

Objectives

1. Fast, honest fit scoring — Give users a clear read on how well a resume matches a job description in seconds, with suggestions they can act on.
2. Tailored outputs — Produce versions of a resume aligned to a specific role without starting from scratch each time.
3. End-to-end job search support — Combine scoring, tailoring, cover letters, and organization (companies, versions) in one product.
4. Sustainable SaaS — Subscription and one-time purchase options, with a clear free tier for discovery.

Features

1. AI resume scoring — Upload a resume and job description for an instant compatibility score and detailed feedback.
2. Resume tailoring — Generate role-specific resume versions optimized for ATS parsing and relevance.
3. Cover letter generation — AI-generated cover letters with tone customization.
4. Rich text editor — WYSIWYG editing with inline AI assistance.
5. Version management — Track multiple versions of resumes over time.
6. Company organization — Group and manage applications by company.
7. Command bar — Spotlight-style search (Cmd+K) for quick navigation across the app.

The Rebuild

The first version was a Next.js App Router app on Supabase with Stripe billing and a Tiptap-based editor. It proved the product out, then came down—the rewrite is a re-architecture rather than a refactor, trading that foundation for a lighter, faster one while keeping the parts that earned their place.

Technology Stack

• Framework: TanStack Start with TanStack Router (file-based routing, generated route tree)
• UI: React 19, TypeScript
• Build & tooling: Vite, Bun, oxlint, oxfmt
• AI: LLM-backed scoring, tailoring, and cover letter generation

Outcome

The original version shipped with AI scoring, paid tailoring and cover-letter flows, and integrated billing. It is no longer running, and the product is being re-architected and rebuilt on a new foundation.