Omelet is the KAIST-backed startup founded in April 2023 that’s building an AI-first foundation for routing and logistics optimization.

OMELET develops a commercial Routing Engine and a Transportation Management System (TMS) under its Optimization-as-a-Service / Infra / System (OaaSIS) offering. The leadership and core engineering team originate from KAIST research groups and include Jinkyoo Park and Changhyun Kwon, supported by specialists in operations research and machine learning.

The engine is designed to solve classical and practical vehicle-routing and pickup-and-delivery problems — TSP, mTSP, CVRP, CVRPTW, preference-aware capacitated routing, and long-running task scenarios — using an AI-first, three-stage workflow: fast encoder–decoder generation of diverse candidate solutions, neural-guided parallel improvement search, and an end-to-end iterative training loop that retrains on internally generated datasets. The team frames this pipeline as an Optimization AI-Agent / an AI foundation model for routing.

OMELET is available as a developer-facing Route Optimization API and as a TMS SaaS platform; you can register on their developer portal, obtain an App Key, and run the TSP/CVRP example payloads in the docs to evaluate solution quality and latency.