Project Description
BACKEND LEAD ENGINEER — FOOD INTELLIGENCE PLATFORM
Node.js · TypeScript · PostgreSQL · Prisma · REST API
Budget: $750 USD · Fixed Price
───────────────────────────────────────
ABOUT THE PROJECT
We are building a Food Intelligence Platform across 7 phases — from data infrastructure through to a consumer grocery app, condition-specific health apps (diabetic, cardiac), genomic personalization, hospital integrations, and an enterprise B2B layer. Phase 1 is backend foundation work. The platform is designed to scale globally across US, EU, Asia, Middle East, and Africa.
This is not a small project. If you deliver Phase 1 well, there is a high probability of continued full-time engagement across all remaining phases. We are looking for someone who wants to grow with the platform, not just complete a one-off task.
───────────────────────────────────────
THE ROLE — BACKEND LEAD (TRUTH LAYER)
You will own the core truth layer of the platform. This means everything that defines how data is structured, interpreted, scored, and kept consistent as the system scales. You define what correct data looks like. You are not moving data — a separate pipeline engineer handles that.
Your ownership covers:
— Full PostgreSQL database schema (20+ tables) with relationships, indexes, and constraints built for long-term scale
— Product identity and deduplication logic — canonical product records, external ID mapping across UPC, USDA, OpenFoodFacts, EU, and Asian databases
— Ingredient knowledge base schema — canonical ingredients with regulatory status by jurisdiction (FDA, EU EFSA, WHO, and others)
— Scoring engine integration — porting and wiring a 7-layer food scoring engine (V3.1) into the backend in TypeScript
— Score auditability — layer breakdowns, engine versioning, ruleset versioning — every score fully traceable
— Condition intelligence — schema for diabetic, cardiac, and other health profiles with evidence-graded clinical rules
— Regional and regulatory infrastructure — Region and RegulatoryJurisdiction tables supporting global operations
— Enterprise isolation — Organization table with data boundary enforcement for hospitals, insurers, and retail partners
— System configuration — FeatureFlag and ScoringCalibrationBenchmark tables
— Reformulation event tracking — detecting and recording when product formulas change
— Medical data stubs — GenomicProfile and ClinicalDataIntegration tables for Phase 4 hospital and genomic integration
— Practitioner referral schema — PractitionerProfile and ReferralEvent tables supporting the referral program
— REST API layer — versioned endpoints (/api/v1/), request/response contracts, barcode lookup, search, scoring endpoints
— End-to-end integration ownership — you prove the full system works correctly together, not just in parts
───────────────────────────────────────
TECH STACK
— Runtime: Node.js with TypeScript
— Database: PostgreSQL (required)
— ORM: Prisma (required)
— API: Express or Fastify
— Queue integration: BullMQ with Redis (you define the schema; a separate engineer wires the workers)
───────────────────────────────────────
PHASE 1 IS COMPLETE WHEN:
— Full Prisma schema implemented — all 20+ tables with correct relationships and indexes
— ProductIdentity deduplication proven — zero duplicate products after bulk import of 1,000+ real products
— Scoring engine wired — all 7 layers running, layer breakdown stored on every score
— Calibration suite passing — benchmark products scoring within expected ranges across all 5 scoring modes
— All API endpoints responding correctly — tested with real data end-to-end
— Feature flags operational
— Organization data isolation proven
— Reformulation detection working
— Engine versioning working — any score reproducible given the same product version and engine version
— Data contract confirmed in writing with the pipeline engineer before independent build begins
───────────────────────────────────────
PAYMENT TERMS
Total: $750 USD — Fixed Price
30% ($225) — Paid upfront at project start
30% ($225) — Paid at midpoint milestone (schema complete, migrations confirmed, scoring engine wired)
40% ($300) — Paid on full completion and acceptance
───────────────────────────────────────
WHAT HAPPENS AFTER PHASE 1
This platform has 7 phases total. Phase 1 is the foundation. Phases 2 through 7 build the intelligence pipeline, consumer apps, enterprise integrations, genomic personalization, and platform layer on top of what you build here. Engineers who deliver Phase 1 correctly are first in line for continued paid engagement across all remaining phases. Given the scale and trajectory of the platform, there is a high probability of long-term full-time retention for the right person.
───────────────────────────────────────
VERIFICATION REQUIREMENTS — MANDATORY
We take verification seriously. The following are not optional:
— Video call interview before any work begins — you must appear on camera and speak to your experience
— Identity verification — valid government-issued ID required
— Portfolio or GitHub review — you must be able to walk through prior backend work live on the call
— English communication — written and spoken clarity is required for this role
— Responsiveness — you must be reachable and communicative throughout the project. Disappearing mid-project is an automatic disqualification from future phases
Applications without a video call will not proceed.
───────────────────────────────────────
HOW TO APPLY
In your application, please include:
1. A brief description of the most complex database schema you have designed — number of tables, relationships, and what it was for
2. Your experience with Prisma and PostgreSQL specifically
3. Whether you have previously ported or integrated a scoring or rules engine into a backend system
4. Your availability and estimated time to complete Phase 1
5. Confirmation that you are available for a video call before work begins
Applications that do not address these points will not be reviewed.
───────────────────────────────────────
Build the truth layer. Everything else depends on it.