ML Platform Engineer

You'll work alongside our CTO and one staff engineer. The Decision Engine ingests forecasts from six external feeds and one proprietary microclimate model, resolves them to a per-consignment risk score, and pushes recommendations to customers' operational systems through a mix of REST, GraphQL, webhooks and direct integrations.

This isn't pure platform work and it isn't pure modelling. The role sits at the intersection — you'll be expected to read a forecast paper one week and a Postgres execution plan the next. You'll talk to customers most weeks. We don't have product managers. Engineers own roadmap.

• —Own the lane-resolution layer of the Decision Engine end-to-end.
• —Improve the cost-modelling subsystem's latency budget (P95 380ms → 150ms target).
• —Build a new integration with one of the major TMS vendors in 2026.
• —Pair with the customer-facing Solutions Engineering team during the first two weeks of every customer onboarding.
• —Be on a one-week-in-six on-call rotation alongside CTO and staff engineer.

• · Five+ years of production backend engineering, with ownership of a service that mattered.
• · Strong Python (or willingness to be productive in Python from day one).
• · Comfort with Postgres at scale, including reading and rewriting query plans.
• · Track record shipping in customer-facing operational software.
• · Direct, low-process working style.

• · Background in aerospace, weather services, freight, or a related domain.
• · Experience with event-driven architectures (Kafka or equivalent).
• · Some applied modelling background (probabilistic forecasting, ensembles).