Sina Bolouri

Project

ApplyGraph

An evidence-first application workspace that maps job requirements to candidate proof, explains fit and gaps, and produces source-linked editable materials.

deployedNext.js 16React 19TypeScriptTailwind CSS 4Supabase/PostgreSQLZodReact Hook FormOpenRouter/Gemini optionalUpstash Redis optionalVercelVitestPlaywright
Visit ProjectGitHub

ApplyGraph

ApplyGraph turns a job description and candidate-provided experience into an explainable requirement-to-evidence graph, fit assessment, gap analysis, and editable application materials. Its score measures how well the supplied evidence covers the extracted requirements; it does not predict interviews, offers, or hiring outcomes.

Problem

Job seekers often tailor resumes and cover letters by matching keywords or generating fluent text without a reliable link back to real experience. That makes it difficult to see which role requirements have strong proof, which are weakly supported, and whether generated claims remain grounded in candidate-provided sources.

Solution

A deterministic pipeline normalizes the job description, extracts requirements, atomizes candidate evidence, ranks up to three matches per requirement, computes weighted coverage, identifies critical gaps, and generates eight source-linked outputs. Match and fit views expose the relationship among requirements, evidence, scores, and gaps. Optional AI may select existing evidence IDs for one summary path, but it cannot add claims or change fit scoring.

User Experience

Users move through job description, resume evidence, projects, and review steps before opening a workbench with mapping, packet, and interactive graph views. They can inspect sources, edit and review generated outputs, resolve gap review states, copy text, and delete the packet. Anonymous mode retains one versioned packet in localStorage. The account interface adds saved packets, profile data, projects, experience, skills, filtering, and pagination when the Supabase schema is connected.

Current Status

Status: deployed. The anonymous production experience at https://applygraph.sinabolouri.com was verified by the source repository with 14 desktop/mobile Playwright tests, and the domain returned HTTPS 200 from Vercel during the 2026-09-22 portfolio review. The repository's current release notes also state that the expected Supabase schema is absent from the production API; hosted signup and saved-account packet workflows are therefore not verified production capabilities yet.

Tech Stack

  • Frontend: Next.js 16 App Router, React 19, TypeScript, Tailwind CSS 4, Radix UI, lucide-react, React Hook Form, next-themes, and locally hosted fonts.
  • Domain: deterministic TypeScript requirement extraction, evidence matching, fit scoring, gap analysis, and output generation with Zod contracts.
  • Backend: Next.js Route Handlers and server-only services.
  • Database and auth: Supabase Auth and PostgreSQL with RLS, restricted RPC writes, graph constraints, and transactional quota enforcement; localStorage for anonymous packets.
  • Optional services: OpenRouter or Gemini for constrained evidence selection and Upstash Redis for shared burst limiting.
  • Quality and hosting: Vitest, PostgreSQL integration tests, Playwright, axe-core, and Vercel.