Skip to main content

Analyzing Edge AI Applications with Arm's CMSIS Debugger for VS Code

Monday, 11 May 2026 Embedded Online Conference 2026, online conference (paid pass) Matthias Hertel (Arm, embedded systems, AI, and IoT for Cortex-M developers)

Arm CMSIS Debugger session, editorial scene of an Indian embedded systems engineer reviewing edge AI inference debug output in Visual Studio Code, with a Cortex-M development board and SEGGER J-Link debug probe on the desk

The Embedded Online Conference 2026 session walks Cortex-M teams through the Arm CMSIS Debugger extension for Visual Studio Code, and how it changes the workflow for understanding, optimizing, and debugging edge AI applications running on Cortex-M devices.

What’s covered

Edge AI workloads on Cortex-M. Edge AI inference on a microcontroller competes with the rest of the firmware for CPU, memory, and timing budget. The session opens with the practical pain points: threading conflicts between the inference task and other RTOS tasks, inference latency that is hard to attribute to a specific layer or kernel, and memory issues that surface only at runtime.

Arm CMSIS Debugger inside VS Code. The talk demonstrates how the Arm CMSIS Debugger extension surfaces the runtime behaviour that traditional debug stops cannot show:

  • RTOS-aware debugging: task lists, scheduler state, and per-thread call stacks for FreeRTOS, RTX5, and other CMSIS-compatible RTOSes
  • Thread inspection: what each task is doing while inference runs, where it is blocked, and which task holds which lock
  • Event tracing: a timeline of scheduler events, ITM/SWO output, and user-defined annotations so the team can see the AI workload’s interaction with the rest of the system

Why this matters for Indian Cortex-M teams. Edge AI inference on Cortex-M55 (Helium / MVE), Cortex-M85, and Ethos-U paired Cortex-M devices is moving fast in Indian wireless IoT, automotive, industrial, and consumer product engineering. The CMSIS Debugger extension is the open-source debug front end that pairs cleanly with the Keil MDK toolchain, Arm Development Studio, and the underlying SEGGER J-Link probe family, so it slots into the toolchain Indian Cortex-M teams already run.

Speaker

Matthias Hertel, Arm

Matthias Hertel

Embedded systems · AI · IoT, Arm

25+ years in embedded systems. Currently focused on democratizing cutting-edge technologies, including AI and IoT, for developers building on Arm Cortex-M.

Session time

RegionTime
US EastMonday 11 May 2026, 13:30 EDT
UKMonday 11 May 2026, 18:30 BST
IndiaMonday 11 May 2026, 23:00 IST

Times sourced from the Embedded Online Conference 2026 schedule. Conference attendance requires an EOC 2026 pass; check the conference page for current pricing and pass options.

Why GSAS is sharing this

GSAS Micro Systems is the authorized India partner for Arm (Keil MDK, Arm Development Studio, ULINK, DSTREAM) and SEGGER (J-Link debug probes). The Arm CMSIS Debugger extension is the open-source VS Code debug front end that talks to those tools, and the edge AI debug workflow it enables is directly applicable to Indian teams shipping Cortex-M55 / M85 / Ethos-U products. If you have questions about how this fits into your India-based Cortex-M program, talk to us after the session.

See the session details on embeddedonlineconference.com →