Food you can trace

// The problem

A food processor keeps its HACCP logs, sanitation checks and lot records on paper, so an audit means hunting through binders and a mock recall takes far too long.

// What we would build

An Android tablet app on the floor for critical control point checks, sanitation and corrective actions, with lot codes scanned at receiving and production, and a web back office that traces any lot one step back and one step forward on demand.

// What is in it

// Stack

• Next.js
• React Native (Android tablets)
• PostgreSQL
• LoRaWAN temperature sensors
• GS1-128 barcode scanning
• AWS

// The problem

A specialty food distributor takes orders from restaurants and grocers by phone, text and email until late at night, and the mistakes show up on the truck.

// What we would build

A web and mobile ordering app where each customer sees its own price list, order guide and delivery days, orders until the cut-off and pays invoices online, with orders flowing straight into the warehouse system.

// What is in it

// Stack

• Next.js
• React Native
• PostgreSQL
• Stripe
• QuickBooks or NetSuite API
• DigitalOcean

// The problem

A beverage and snack maker reformulates often, and each change means a regulatory specialist checks label artwork, ingredients and allergen statements against the recipe by hand.

// What we would build

An agent that reads the label artwork and the recipe, checks ingredient order, allergen statements, bilingual text and nutrition facts against the rules your team has written down, and lists each issue with the rule it comes from for the specialist to decide. It runs on a frontier model through its API, which reads artwork well and costs little at a few labels a week; if your recipes are a closely held secret, an open-source model in your own cloud is the alternative.

// What is in it

// Stack

• Frontier model API with vision
• Python
• Next.js
• PostgreSQL
• DigitalOcean

// The problem

When a load of produce or seafood arrives warm, the packer, the carrier and the retailer each hold their own temperature data, and the claim turns into an argument over whose logger to believe.

// What we would build

A private permissioned ledger where packer, carrier and retailer each run a node, recording trailer and pallet logger readings, door openings and hand-offs as they happen, so a disputed load is settled from one record nobody could change after the fact. If the others already trust a database run by one of them, that is cheaper, and we will say so.

// What is in it

// Stack

• Hyperledger Fabric (private, permissioned)
• Node.js
• Next.js
• Bluetooth temperature loggers
• PostgreSQL
• AWS

// The problem

A food producer receives purchase orders from grocery chains by EDI through a service that prints them, and staff type each order, advance ship notice and invoice into the ERP by hand.

// What we would build

An integration that receives orders from each retailer, creates them in your ERP, sends back acknowledgements, ship notices and invoices in the format each retailer requires, and alerts your team to any document that is rejected.

// What is in it

// Stack

• Node.js
• PostgreSQL
• EDI X12
• OpenAPI
• DigitalOcean

// The problem

A food plant knows its yield and waste only at month end, from production sheets and ERP exports pieced together by the plant manager.

// What we would build

A dashboard loaded every shift from your ERP, scales and production records, showing yield, giveaway and waste by line, product and shift, so supervisors see a drift the same day rather than at month end.

// What is in it

// Stack

• Next.js
• PostgreSQL
• dbt
• Metabase
• DigitalOcean