advin.io
A boutique AI research lab in Dallas, TX. Handcrafted systems for industrial intelligence — spec data, robot emulators, device fleets — built in the metroplex, with deployed models running live on Replicate.
Cross-vendor industrial spec data — motors, drives, gearheads, contactors, robot arms — indexed, filtered, and exportable. No marketing copy on the rows, no request-a-quote gates.
Spec-faithful emulators of autonomous-mobile-robot REST APIs — MiR today, more brands coming. Develop and test fleet software without a robot on the floor.
Fleet operations for connected hardware, all in Rust — per-device mTLS identity, a live dashboard, firmware tracking. From ESP32 register to cloud database in one language.
A microserver template in C++ — the small, sharp starting point for services that have to fit where full frameworks don't.
Bootstrap efficient machine-learning models — tooling for getting small, fast models into production shape.
A worked example of serverless CPU inference — an endpoint pattern for running models without keeping a GPU warm.
An endpoint that extracts COLMAP photogrammetry data from a plain video — 3D reconstruction as a service call.