2026 GSoC Ideas#

Ideas#

Web BeagleBoard Imager Medium complexity 350 hours

Web UI for BeagleBoard Imaging Utility

Allow flashing BeagleBoard.org boards directly from web based UI, without requiring any local install. BeagleBoard Imaging Utility is written using iced, which does allow targeting web. Depending on the performance, using Tauri/Dioxus or something similar for UI parts is acceptable. However, the flashing code should be shared with the existing imager.

Goal: Working BeagleConnect Freedom flashing
Stretch Goal: Working SD Card flashing
Software Skills: Rust, Webassembly
Possible Mentors: Ayush Singh

Enhanced HDMI & Splash Support for BeaglePlay in U-Boot Medium complexity 350 hours

Enhanced HDMI & Splash Support for BeaglePlay in U-Boot

This project brings HDMI output to U-Boot on BeaglePlay by adding support for the IT66121 DPI-to-HDMI bridge and improving U-Boot’s splash screen system. The goal is to make the board give clear visual feedback as soon as it powers on.

With this, BeaglePlay will be able to show a splash screen on HDMI during boot, complete with proper scaling, rotation, and optional progress indicators. This makes the boot process easier to understand and much more friendly for headless setups.

Goal: Enable HDMI display output during U-Boot on BeaglePlay with a modern, configurable splash screen system, suitable for upstream.
Hardware Skills: Basic understanding of display interfaces (DPI, HDMI), I2C
Software Skills: U-Boot internals, device tree, embedded C, Linux display pipelines
Possible Mentors: Dhruva gole, Devarsh Thakkar (tentative)

Implement Missing Greybus Protocols Medium complexity 175 hours

Add implementations for Greybus protocols missing in Greybus for Zephyr. The priority protocols are Camera and Audio. Optionally, SDIO, HID, USB and Power Supply protocols can also be implemented, if time permits.

The tested would be performed using BeaglePlay and BeagleConnect Freedom. Addionally, tests running on native simulator should also be added for any new protocols.

Goal: Implement Camera and Audio greybus protocols.
Stretch Goal: SDIO, HID, USB and Power Supply protocols.
Hardware Skills: Basic understanding of UART, 6lowpan.
Software Skills: ZephyrRTOS, TCP, C.
Possible Mentors: Ayush Singh

BeagleMind Model Fine-Tuning for Offline Troubleshooting Medium complexity 175 hours

Fine-tune a large language model (LLM) on curated question-answer pairs for BeagleBoard troubleshooting. The project aims to enable contributors and users to resolve issues offline, using an efficient, locally running model.

Goal: Empower users to troubleshoot BeagleBoard hardware offline with a custom QA model
Stretch Goal: Apply quantization to the QA model so it can run efficiently on embedded and resource-constrained devices, enabling offline troubleshooting even on low-power hardware.
Software Skills: Python, supervised fine-tuning, LLMs, data preprocessing, model evaluation, deployment, PyTorch/TensorFlow
Hardware Skills: Quantization techniques, resource-constrained inference, hardware acceleration (optional)
Possible Mentors: Jason Kridner, Fayez Zouari

New ideas#

Visit our forum to see ideas being discussed!