AI-Powered Recruitment Marketplace (Next.js + AI Document Processing + Visa Pipeline)

QUESTRA RECRUITMENT — AI-Powered Recruitment Marketplace (Next.js + AI Document Processing + Visa Pipeline) PROJECT OVERVIEW We're building a recruitment marketplace that connects skilled tradespeople working overseas with employers in Australia, combined with an AI-powered document processing pipeline and an integrated visa/immigration case management system. The platform is operated by a registered Australian migration agency, so data security, privacy compliance, and accuracy are not optional — they are core requirements. This is a genuinely interesting build: a three-sided marketplace (candidates, employers, internal admin) with an LLM-powered document pipeline at its core, not a template CRUD app. We're looking for a developer or small team who has actually shipped something with comparable complexity — not just "can do it" but "has done it." We have a complete, detailed specification document (11 pages, covering every module, user role, and requirement) ready to share with serious applicants. This posting is a summary — the full brief is what you'll be quoted and held against. WHAT WE'RE BUILDING Three connected portals: 1. Candidate portal — mobile-first intake flow where users upload a CV, ID document, and qualifications, which get processed by an AI pipeline (extraction, reformatting, sensitive-data redaction) before being reviewed by our internal team. 2. Employer portal — desktop-first browsing experience where employer accounts view AI-processed, redacted candidate profiles, save favourites, and request to unlock full contact details. 3. Admin dashboard — internal review queue where our team approves AI output, edits fields, manages the case pipeline, and tracks progress through multiple sequential approval/compliance stages. Plus an AI processing pipeline that: - Extracts structured data from uploaded CVs/documents (OCR + LLM extraction) - Rewrites and reformats content to a target market's professional standard - Maps job titles to a standardised occupation code list (mapping table will be supplied) - Automatically redacts personally identifying information (surname, phone, email) before any document is employer-visible - Generates a confidence score for human reviewers CORE REQUIREMENTS — READ BEFORE APPLYING This is the single most important rule in the entire build: an employer account must NEVER be able to access a candidate's unredacted contact details or surname under any circumstance, until that candidate has explicitly accepted contact. This needs to be enforced at the DATABASE level (row-level security), not just hidden in the UI. If you don't have direct experience implementing row-level security in Postgres/Supabase, please don't apply — this is the one area we will not compromise on. Beyond that: - All candidate documents (passports/IDs, CVs, certificates) must be stored encrypted, in access-tiered cloud storage - Full audit logging of every access to a candidate's personal documents - The build needs to be consistent with general data privacy best practice (we are Australian-based and subject to the Australian Privacy Principles) — prior experience building for a regulated or privacy-sensitive industry (health tech, fintech, legal/immigration tech) is a strong plus TECH STACK (preferred, open to your recommendation if justified) - Frontend/backend: Next.js 14+ (App Router) - Database/auth: Supabase (Postgres + Auth + Row-Level Security) - Document storage: AWS S3 with KMS encryption - AI/LLM: Anthropic Claude API for extraction, rewriting, and classification - OCR: AWS Textract or equivalent - Hosting: Vercel - Payments: Stripe (subscription billing) - Async processing: Upstash Redis or equivalent job queue We're flexible on stack if you have strong reasons for an alternative — just explain your reasoning in your proposal. INTEGRATIONS REQUIRED - WhatsApp Business API (via 360dialog or Twilio) for automated candidate communications - GoHighLevel (lead capture / marketing automation — we manage this, you just need to integrate via webhook/API) - A recruitment CRM (API/webhook sync for pipeline tracking) - Stripe (employer subscription billing) - An e-signature provider (SignWell or DocuSign) for client agreements WHAT WE NEED FROM YOU - Confirmed experience with Postgres/Supabase row-level security — please describe a specific project where you implemented this - Confirmed experience integrating an LLM API (Claude, GPT, etc.) into a production application for document processing, data extraction, or similar — not a chatbot widget - Next.js + TypeScript production experience (please link 2-3 live examples) - Comfortable working to a detailed written specification (we will share the full 11-page brief with serious applicants) - Available for regular video calls during Australian business hours (AEST) for at least the first 2-3 weeks ENGAGEMENT STRUCTURE We're planning a phased build: - Phase 1 (MVP): core intake flow, document upload, integrations, basic employer browsing — target 6 weeks - Phase 2 (Full platform): complete AI pipeline, full admin dashboard, full employer marketplace, security hardening — target 10 weeks following Phase 1 - Phase 3 (ongoing): ongoing feature development, open to retainer or sprint-based arrangement post-launch Open to: fixed-price per phase, or capped time & materials with milestone payments. Please indicate which model you prefer and why. HOW TO APPLY Please include in your proposal: 1. A specific example of a project where you implemented row-level security or similarly strict access control 2. A specific example of LLM/AI integration into a production app (not a demo or hackathon project) 3. Your proposed team composition (named individuals if an agency, not just headcount) 4. Your initial thoughts on timeline and rough budget range for Phase 1 + Phase 2 combined 5. Links to 2-3 live, comparable projects Generic proposals that don't address points 1 and 2 directly will not be considered — this is the part of the project that actually matters most, so please don't skip it. We're looking to move quickly with the right partner. Looking forward to your proposal.