Hi,
I am looking to hire a full-stack and/or back-end engineer with experience in eCommerce/AdTech (Product and Shopping Feeds) that would work alongside a designer (Tailwind) and Shopify Developer to build what Iād like to call a universal offer router.
This contract would be for the entire software development lifecycle, deployment, and maintenance of the product.
Technical Stack: Modular Monolith/SPA
(Minimizing the "network tax" and complexity of microservices)
ā¢Backend: Go (Chi or Eco)
⢠Frontend: HTMX + Templ + Tailwind
⢠Database: PostgreSQL+ Vector Database (Pinecone/Milvus?) for product matching
⢠Job Que: Redis/Asynq
⢠Auth: JWT via Magic Link
⢠API: GraphQL (Shopify-Native)
⢠Hosting: AWS/GCP/Azure (or Railway?)
⢠??Additional Infrastructure??: Kafka, Spark, and Cassandra/ScyllaDB
Shopify Application: Shopify Native (hired separately)
⢠React + Remix
⢠Polaris Web Components
⢠API: GraphQL (endpoint from monolith)
⢠Auth: Shopify Auth and Billing
⢠Draft Order API
WIX: TBD (hired separately)
Software Architecture Overview:
While any engineer could be qualified to build this system, an individual with domain experience who understands the business logic behind Catalog Ads (DPAs), Merchant Center, Product/Shopping Feed API Ingestion will understand why the product/system is intended to be built this way.
I am specifically looking for what the industry would be call āa feed guyā (or girl). While a state machine is critical, this is not a real-time bidding system/auction. Here are some roles that Facebook, Google, Pinterest, etc would call this role:
Commerce Manager/Platform API Engineer, Merchant Center Engineer, Ads Content API Engineer, Catalog Ads Infrastructure Engineer, DPA API Engineer, Product Feed Engineer, Product Feed Manager, Ads Infrastructure Engineer, Shopping Feed Engineer
A. The Ingestion & Intelligence Plane (The "Brain")
⢠URL Normalizer: Experience with regex-based cleaning, handling tracking parameters (UTMs), and resolving redirects to ensure unique product identification.
⢠Vector/Fuzzy Matching Engine: Knowledge of fuzzy matching algorithms or using vector embeddings (Pinecone, Milvus) to identify that "Nike Air Max 90" from Vendor A is the same as "AM90 Shoe" from Vendor B against an Imported Master Catalog to identify the product and its MSRP.
⢠State Machine (The Negotiation): Managed by a workflow engine like Temporal. It handles the back-and-forth "Counter-Offer" logic, maintaining the status (Submitted, Countered, Accepted, Expired) for every offer ID.
B. The Integration Layer (The "Bridges")
⢠SME Bridge (Shopify/Wix/Square): A native app that uses the Draft Order API to generate checkouts and the App Billing API to automatically collect the 14% fee via the merchantās existing platform invoice.
⢠Enterprise Bridge (Nike/etc): A Server-to-Server (S2S) API connection. It uses an Attribution ID passed through the URL. Nikeās backend confirms the sale via a webhook or API callback, and the 14% fee is settled via a monthly B2B invoice.
C. The Identity & Privacy Layer (The "Ghost")
⢠Auth: Passwordless OTP (One-Time Password) via SMS/Email.
⢠Storage: Verified identities are stored as Hashed Strings. No PII is stored in the clear.
⢠Session Management: Uses JWTs (JSON Web Tokens) stored in the browser's LocalStorage, triggered by Magic Links sent to the user. This allows a "Wish List" experience without cookies.
D. The Analytics Layer (The "Ledger")
⢠Immutable Savings Ledger: A transaction log that records Original MSRP - Final Settled Price only when a PAID signal is received from the merchant API.
⢠Time-Series Aggregation: Uses Materialized Views to provide the user with real-time "Annual" and "Lifetime" savings stats on their dashboard.
The role would start as a paid consultation that could move into full-time development.
This job is not time sensitive and I'd be happy to wait for the best candidate you are busy.
Being located in the Bay Area would be preferred, but not required.
Minimum of 6+ years experience.