How the Rust Project Selected Its Google Summer of Code 2026 Projects: A Step-by-Step Guide

Introduction

Every year, Google Summer of Code (GSoC) invites open-source projects to mentor new contributors. The Rust Project, a regular participant, recently went through its own selection process for GSoC 2026—and came out with 13 accepted proposals. This guide walks you through each stage of that process, from publishing project ideas to celebrating the final list. Whether you're an open-source mentor, a prospective contributor, or just curious about how large projects manage GSoC, these steps reveal the behind-the-scenes work that led to a successful cohort.

What You Need

• A clear understanding of the GSoC program (timeline, rules, obligations).
• A set of project ideas that align with your organization's priorities.
• A communication platform (e.g., Zulip, Discord, mailing list) to discuss ideas with applicants.
• Mentors with enough bandwidth to guide contributors—and backup plans in case funding changes.
• An evaluation rubric to rank proposals fairly (quality, community importance, prior contributions, etc.).
• Access to proposal submissions from applicants (typically through the GSoC portal).
• Patience to handle AI-generated content and low‑quality contributions.

Step-by-Step Selection Process

Step 1: Publish Project Ideas and Open Discussions

Months before the application deadline, the Rust Project published a list of GSoC project ideas on its website. These topics were chosen to benefit both the project and the wider Rust community. To attract good candidates, the team started discussing these ideas on Zulip, answering questions and encouraging early engagement.

Step 2: Engage with Potential Applicants

Many applicants began participating in the Zulip conversations well before GSoC officially started. Some even made non‑trivial contributions to Rust repositories during this pre‑application phase. This early interaction allowed mentors to gauge an applicant's communication skills, motivation, and technical level.

Step 3: Receive and Review Proposals

By the end of March, the Rust Project received 96 proposals—a 50% increase over the previous year. While the team welcomed the high interest, they also noticed a growing challenge: AI‑generated proposals and low‑quality contributions produced by AI agents. Although these made the review process more difficult, the Rust Project found it manageable by focusing on genuine interactions and substantive contributions.

Step 4: Evaluate Proposals Thoroughly

Mentors examined each proposal against multiple criteria:

• Prior interactions with the applicant (from Zulip, GitHub, etc.).
• Contributions so far (patches, bug fixes, documentation improvements).
• Quality of the proposal (clarity, feasibility, timeline).
• Importance of the proposed project for Rust and its community.
• Mentor bandwidth and availability—a critical resource.

Because every project topic received multiple proposals, the team had to select only one per topic. They also avoided assigning multiple projects to a single mentor to prevent overload.

Step 5: Handle Constraints and Cancellations

Not all planned projects could be supported. Several mentors lost their funding for Rust work in the past few weeks, forcing the Rust Project to cancel some project ideas. Despite these setbacks, the available mentor pool was sufficient to support the strongest proposals.

Step 6: Rank and Submit the Ordered List

GSoC requires each organization to submit an ordered list of its best proposals. The Rust Project narrowed the original 96 down to the proposals that could realistically be mentored with their available resources. This ordering is always challenging because Rust is a large project with many priorities, but the team balanced all factors and submitted the list to Google.

Step 7: Google Announces Accepted Projects

On April 30, Google revealed the accepted projects for GSoC 2026. The Rust Project was thrilled to learn that 13 of their proposals had been accepted—a record number for them. The final list (in alphabetical order) included:

• A Frontend for Safe GPU Offloading in Rust – Marcelo Domínguez, mentored by Manuel Drehwald
• Adding WebAssembly Linking Support to Wild – Kei Akiyama, mentored by David Lattimore
• Bringing autodiff and offload into Rust CI – Shota Sugano, mentored by Manuel Drehwald
• Debugger for Miri – Mohamed Ali Mohamed, mentored by Oli Scherer
• Implementing impl and mut restrictions – Ryosuke Yamano, mentored by Jacob Pratt and Urgau
• Improving Ergonomics and Safety of serialport-rs – Tanmay, mentored by Christian Meusel

Step 8: Celebrate and Prepare for the Coding Phase

With the accepted projects announced, the Rust Project now enters the active mentoring phase. Contributors begin coding, mentors provide guidance, and the community watches the new features take shape. This step is as much about building relationships as it is about writing code.

Tips for Running a Smooth GSoC Selection

• Start early. Publish project ideas and open discussions at least two months before the deadline. The earlier you engage, the more authentic the proposals will be.
• Screen for genuine interest. Look for applicants who ask thoughtful questions and make small contributions before applying. These are less likely to be AI‑generated.
• Plan for AI proposals. Use your evaluation rubric to favor concrete details and community involvement. Ask applicants to link to prior work in your repositories.
• Be realistic about mentor availability. A mentor can only handle one or two projects effectively. Factor in potential funding losses when drafting your project list.
• Communicate clearly. Let applicants know the criteria early (e.g., importance of the project, prior contributions). Transparency reduces last‑minute surprises.
• Don't be afraid to say no. Even with 96 proposals, quality matters more than quantity. Focus on the proposals that align with your project's roadmap and mentor capacity.

For more details on the Rust Project's GSoC journey, revisit Step 1 to understand how the ideas were shaped, or jump to Step 3 for insights on handling AI‑generated proposals. Good luck with your own selection process!