ROS 2 learning path v0.1 reveals zero competency models in 14 programs

In a detailed discussion post on July 21, 2026, community member NemanjaGalic released v0.1 of a proposed ROS 2 learning path architecture. After benchmarking 14 robotics programs (including Northwestern and TU Delft) from primary sources, he found that none define a competency model, offer branched progression, or provide open criterion-referenced assessment. All current curricula are linear sequences (Phase 1→5, Week 0→6), lacking the layer that describes what a ROS 2 developer can demonstrably do at each stage independent of any course. The retreat of universities from open teaching (partly due to grant expiration) leaves a structural hole that Galic argues must be filled by a community-maintained model, similar to how ROS documentation itself is sustained.

To bridge this gap, Galic designed a four-stage competency ladder backward from employability: Explorer (can read a running robot), Builder (can make one robot do one thing), Integrator (can assemble and debug a system), and Production developer (the employability threshold). Each outcome is a verifiable “can …” statement rather than course completion. The job market backs this need: a sample of 11 postings (7 from /c/jobs, 4 from Indeed) showed 10/11 ask for production-level ROS 2, 8/11 for C++ and Python, and 5/11 for navigation/SLAM. The architecture is explicitly a discussion draft — not a course or product — and invites community feedback to shape a lasting learning resource.