Forklifts, Facts, and Functions: Building a Warehouse Management System with Clojure+Datomic - Pote

Tim Pote discusses the architecture of a custom warehouse management system built with Clojure and Datomic, highlighting state machine modeling and time-travel features.

When we were asked to bring order to the chaos of a 100,000 sq ft warehouse run on spreadsheets, the default advice was clear: “just buy something.” But after seeing the brittle integrations and inflexible workflows of off-the-shelf systems, we took the road less purchased. In this talk, I’ll walk through the architecture of the Warehouse Management System (WMS) we built in Clojure and Datomic. I’ll cover: * How we got the buy-in to build * How we modeled the entire warehouse as a state machine * What happened when financial tracking interfered with workflows on the floor * How Datomic’s time traveling saved us after an accidental reset * Datomic Cloud gotchas and how we mitigated them You'll walk away with concrete patterns for building resilient systems in Datomic—plus a few hard-earned lessons on what to avoid when your ops team also wears steel-toed boots. Biography Tim Pote is a backend engineer and consultant with 14+ years of experience building distributed systems. He has deep expertise in Clojure and Datomic and authored Fusebox, a fault-tolerance library for Clojure. Most recently, he led the design and development of a warehouse management system from scratch—turning forklifts and spreadsheets into a state machine. Tim runs Edgedancer, a consultancy focused on helping Clojure companies build systems that scale with ease. Recorded Nov 13, 2025 at Clojure/Conj 2025 in Charlotte, NC.