Senior Full-Stack Engineer — Build a Production Property-Records Search Platform

Senior Full-Stack Engineer — Build a Production Property-Records Search Platform (NYC Open Data → Verified Reports) Engagement: Project-based with a clear MVP, then ongoing phases for the right person Type: Long-term potential · Expert level only Location: Remote (US East Coast hours preferred for overlap) Who we are We're an established New York real-estate services company. We've spent two decades doing title and settlement work, and over the last year we've built a serious internal automation stack — AI pipelines, CRM intelligence, financial systems, and a working data engine that already pulls property records from NYC's open-data ecosystem. That engine is the foundation. We now need a senior engineer to wrap it in a real production platform that our team uses every day to order, run, verify, and deliver property searches — replacing the outside vendors we currently pay for the same data. This is not a "spin up a quick CRUD app" job. It's a real systems build with messy government data, certification/liability implications, and a daily-use internal product at the end of it. We're looking for one strong person (or a tight lead + helper) who can own the whole thing. The project in plain terms Today we pay third-party vendors to run municipal and title searches on properties. The underlying data is almost entirely available for free from official NYC and NY State sources. We've already proven we can pull it. What we need is the production layer around that data: An ordering + workflow portal our production team works in — order a search against a property and file, watch its status, review results. A data + verification layer that queries 80+ official government datasets, reconciles inconsistencies, and attaches a live source link to every single result line so any item can be verified against its origin in one click. A delivery layer that generates clean, branded PDF reports and tracks everything by internal file number, with the ability to write results back into our existing title-production software via its API. Think of it as building the product that turns raw open data into a certified, client-ready search — the thing our vendors charge us for, brought in-house. What you'll build (scope) Phase 1 — MVP (the priority): A property-records search portal for NYC (five boroughs) covering municipal/departmental search types (certificates of occupancy, violations across multiple agency systems, tax/water/sewer, fire, housing, zoning, environmental, etc.). A resilient data pipeline querying official open-data APIs (Socrata/SODA, ArcGIS REST, address/parcel geocoding, agency record systems) and reconciling them into a single normalized result set per property. Verify-link on every result line — each item links back to its official source record so a human can confirm it in one click. This is non-negotiable and central to the product's value. Address → parcel resolution (NYC borough/block/lot) as the spine that ties every dataset together. A status dashboard: searches grouped under files, with clear green / needs-review / problem states. A human-in-the-loop review step before anything is finalized (confidence scoring + a sign-off gate). Branded PDF report generation. Everything tracked and retrievable by internal file number. Later phases (for the right long-term partner): A title-chain / public-records module — recorded documents, judgments, liens, lis pendens, tax warrants, estate/probate records, owner-name searches across the city. Write-back integration into our existing title-production platform via its API. Approval/notification workflow integrations (e.g. Slack-based sign-off). Expansion beyond NYC to surrounding counties, which work on a fundamentally different data model (no unified open-data API — more ordering/intake automation than direct pulls). We'll share the full internal spec, the existing engine, and our data-source map with the finalist. Do not propose a from-scratch rebuild that ignores what we already have — a big part of why this is efficient is that the hardest data work is partly done. Tech stack We've standardized on a Cloudflare-native serverless stack for our newer systems: Cloudflare Workers (compute), D1 (database), R2 (object/PDF storage), Queues (async jobs), Cron Triggers (scheduled pulls). Modern JS/TS frontend (React or similar) for the portal UI. Dynamic PDF generation. REST/JSON API integrations and webhooks throughout. If you're elite on an adjacent serverless stack (AWS Lambda/DynamoDB/S3, etc.) and can justify it, we'll listen — but Cloudflare experience is a strong plus and keeps us consistent with the rest of our infrastructure. Required experience (please don't apply without these) Production serverless / edge development — you've shipped real systems on Cloudflare Workers (strongly preferred) or a comparable serverless stack, not just toy projects. Government / public-records / open-data integration — you've worked with Socrata/SODA APIs, ArcGIS REST services, or similar official data sources, and you understand that government data is inconsistent, paginated, rate-limited, and occasionally just wrong. You know how to build pipelines that stay reliable anyway. Data normalization & reconciliation — taking many messy sources and producing one clean, trustworthy result set, with provenance preserved for every field. Full-stack delivery — you can build the backend pipeline and a clean, usable internal portal UI. Geospatial / address resolution — geocoding, parcel/lot identification, joining datasets on a property key. Dynamic, production-quality PDF generation. Third-party API integration — consuming and integrating REST APIs, handling auth, webhooks, retries, and failure gracefully. Strong bonus points Domain experience in title insurance, real estate, proptech, legal-tech, or fintech — anyone who has built property-records, public-records, or compliance-data tooling will be far ahead. Experience with human-in-the-loop / QA-gated workflows where outputs carry liability and can't just be auto-published. Familiarity with NYC-specific systems (ACRIS, DOB/HPD/ECB record systems, PLUTO, NYC Open Data) is a major edge. Experience integrating with closing/title-production software APIs. How to apply Generic proposals will be ignored. To show you actually read this, please include: A specific example of a system you built that pulled from messy external / government / third-party data and made it reliable. What broke, and how you handled it. Your experience with Cloudflare Workers / serverless (or your honest case for an alternative stack). How you'd approach attaching a verifiable source link to every result line at scale — your high-level thinking, not a full design. Whether you've touched real-estate, title, public-records, or compliance data before. Your availability, your working hours / timezone, and how you prefer to scope and price a build like this (fixed-phase, milestone, hourly — tell us how you work best). We're looking for a long-term technical partner, not the cheapest bid. If you're the right person, Phase 1 is the start of a multi-phase relationship. Screening questions (for the Upwork application form) Copy these into the "Additional questions" section when posting: Describe a production system you built that integrated unreliable external or government data. What was the hardest part? Have you shipped on Cloudflare Workers/D1/R2/Queues? If not, what serverless stack do you use and why? Have you ever built tooling for real estate, title, public records, or compliance? Briefly describe. How would you keep a data pipeline that depends on 80+ external APIs from silently breaking? What's your preferred way to scope and price a phased build like this one?