Full-Stack Developer to Help Drive a Multi-Module SaaS Platform to MVP

About the platform Integrated platform for labour relations — built for the union and employer-side teams who negotiate collective agreements, handle grievances, and represent members. Bargaining and dispute resolution carry an enormous amount of administrative overhead: documents scattered across formats, costing done in fragile spreadsheets, deadlines tracked by hand, no shared record of what was proposed when. The premise of the platform is that smart tooling and real integration between these workflows can take most of that burden off the table, so the people doing the work can focus on the substance — the actual disagreements, and getting to a deal or a resolution. It's a working system built across several connected modules, all sharing one set of records about organizations, bargaining units, and agreements. I've built it solo so far — one person working with AI tooling. It's reached the point where I need a second engineer. The modules Roughly in order of where I most need help: Agreements repository — the authoritative store for collective agreements and related documents. • Ingesting and parsing uploaded documents into structured, clause-level records • Semantic search across agreements by meaning rather than keywords • Comparators across agreements, and extraction of wage and rate data for use elsewhere in the platform Engineering: document parsing and structured extraction, embeddings and vector search, LLM pipeline orchestration, and data modeling for deeply nested, versioned documents. Grievance management — running a grievance from intake through to resolution or arbitration. • Tracking procedural deadlines that chain off when steps actually happened, including extensions, holds, and hearings • AI-assisted intake and conversational interviews with grievors and witnesses, always human-reviewed • Correspondence and service-of-documents exchange with the employer, including an Outlook add-in • Evidence management, brief assembly, and an immutable audit trail Engineering: correctness-critical rules and date computation, LLM integration with guardrails, and email/third-party integration. Member surveys — gathering member priorities to build a bargaining mandate. • Survey authoring with branching logic, conditional paths, and many question formats • Statistical analysis turning member responses into ranked, weighted bargaining priorities • AI follow-up probing that drills into a respondent's own answers to get past shallow responses Engineering: complex interactive UI state, integrating real statistical computing into the application, and streaming conversational AI flows. Bargaining workflow — what happens at the table. • Proposal exchange clause by clause, with turn-taking between the two sides • Staging proposals into a round, sharing it across the table, and keeping a full record of how each clause evolved • Clause-level editing with attributed tracked changes Engineering: workflow and state-machine logic, rich-text editing internals, diffing and version history, and multi-user coordination. Costing and scenario modeling — what proposals actually cost. • Costing proposals across a multi-year contract term, with compounding • Wage grids and side-by-side scenario comparison • Custom data modeling for the operational realities of specific industries • Producing the mandate and bargaining-book documents a team walks into the room with Engineering: calculation-engine correctness, industry-specific modeling engines, data-dense UI, and document generation. Compliance checking — testing agreement language against the law. • Legislation and regulations encoded as structured, versioned, citable rules • Automated checking of agreements against those rules • Document quality review — ambiguity, inconsistency, drafting problems • A review queue where findings get triaged and dispositioned Engineering: rules-engine design, matching document language to applicable rules, and review workflow. Platform layer — what everything else runs on. • Access and licensing across organizations and bargaining units • User and organization administration • Stripe subscription billing • In-app bug and feedback tracking Engineering: access-control and permission resolution, payments and subscription billing, multi-tenant data modeling, and administration interfaces. The work You'd start with what's most immediately useful: fixing what's broken, working through the flows heading toward an upcoming stakeholder demo, and taking on what I hand you as priorities move. There's substantial work here — I'll be setting direction and reviewing, especially early on, but this isn't a queue of trivial tickets. Past that push, there's real potential for this to continue as an ongoing engagement — bug fixes, stabilization, and new feature development. A great deal of this codebase was written with AI assistance, working against detailed written conventions. You're welcome to work the same way or not — I care about the output and whether it fits the spec and works, not how you got there. What I do want is someone who'll push back. Fresh eyes that challenge how things are built are a large part of why I'm hiring, and I'd rather hear the objection than have it quietly worked around. How we'd work GitHub, pull requests, I review. Regular check-ins, and I'll make myself available to unblock you — particularly in the first few weeks while you're finding your footing in a large codebase. Stack TypeScript throughout — Node.js and Express on the server, React and Vite on the client, PostgreSQL with Drizzle ORM. TanStack Query, Tailwind, Radix. Anthropic APIs for the AI features, pgvector for search, R for the statistical work. Who I'm looking for The person who does well here is comfortable landing in a large system they didn't build, can read existing code and conventions and work within them, and asks when something's ambiguous rather than guessing. Someone who writes tests for what they touch, and who'll tell me when they think I've got something wrong. Must have: • Strong full-stack TypeScript experience — React and Node in production • Real experience working in large existing codebases, not only greenfield projects • 20–30 hours a week available, with a few hours of daily overlap with Eastern time • Able to start within the next month Nice to have: • Document parsing, semantic search, or production LLM integration • Rules-engine or calculation-heavy work where correctness really matters • Background in labour relations, legal tech, or HR software — a genuine plus, not required • R or statistical computing, for the survey side To apply Please include: • Links to real work — shipped products, anything you can point to • A short story about working in someone else's large codebase: what you were dropped into, and how you got your bearings • One specific reaction to something in this posting — a question, a doubt, something you'd want to dig into. It doesn't need to be flattering.