Radar

Finding a job is broken not because jobs don’t exist, but because the system was built for employers, not candidates. Companies use automated systems to filter thousands of applicants in seconds while job seekers respond with manual effort: scrolling across a dozen platforms, rewriting CVs for every role, filling in the same forms repeatedly, and never knowing why they were rejected. The best opportunities disappear within hours of posting, yet most candidates apply too late with CVs too generic to survive an ATS filter. Radar fixes this asymmetry. It monitors every major French job platform in real time and the moment a matching role appears, it fires an instant alert with three things already done: a CV tailored to that specific job with a verified ATS improvement score, a personalised cover letter that references the company’s actual context, and a tracked entry in your application pipeline. A conversational intelligence agent extracts the quantified achievements your CV never captured, and a hiring manager discovery tool identifies and drafts a personalised outreach message to the person on the other side of the decision so your application never goes into a black hole unacknowledged.

Every outcome Radar tracks feeds back into the system. Which CV versions got interviews. Which companies responded. Which outreach messages worked. Over time Radar builds a precise picture of what actually gets you to interview not generic statistics, but personalised conversion intelligence that improves with every application sent. . Radar gives every candidate the same automated advantage that employers already have: the intelligence to not just apply more, but to convert more.

During the first two sprints the team built the architecture for multiple users: Supabase Auth with Google OAuth, row-level security across all tables, a full onboarding flow, and user-isolated data throughout. The frontend was refactored from a monolith into a structured component and feature directory with a two-layer semantic design token system, a production-grade dark UI, and a responsive landing page.

The New product features built this sprint include automated background job polling via Supabase Edge Functions, real-time push notifications when a match is found, an ATS before-and-after scoring system shown inside the CV modal, a pitch generation feature alongside CV generation, a pending approvals panel for reviewing auto-generated CVs before they are saved, Stripe billing with Free and Pro tier gating, and an automation settings panel for configuring keywords, match thresholds, and check frequency. The conversational intelligence agent powered by MuleRun, hiring manager discovery via Claude-generated Boolean search, cover letter generation, and the ATS quality gate shown before submission were all designed and scoped during this sprint and are either in active development or demo-ready as of this submission.