@geiger_j atleast reply something…
Dear Mukul,
Thank you for sharing your perspective. I want to address several of the points you raise directly.
On the timing of personal feedback: the public announcement only went out this Thursday, April 30. Follow-ups to the candidates we interviewed, who dedicated time for the video calls on short notice, are on our list for the coming days. For anyone else who would like personal feedback on their application, you are of course welcome to reach out to me directly via DM here on Discourse.
On the process and what we communicated: we did give applicants advance notice. In my latest post on this thread, before the official announcement, I explicitly wrote that anyone who did not receive an interview invitation had not been selected for our final ranking, that we had received many applications, and that we could only interview a small number of candidates. That was meant precisely so that people who were not invited to an interview would not be left guessing until April 30. To be clear as well: if you received a generic rejection email, that was not sent by us directly. Those notifications go out through the GSoC / NumFOCUS pipeline, and we have no control over their wording.
There are also restrictions on what we can communicate during, and after, the selection process, both from NumFOCUS and from Google, that limit how much detail we are able to share about individual rankings. We have to operate within those guardrails. Beyond that, AiiDA was allocated a single slot for a project with a lot of qualified applicants, which by definition means that the majority of proposals will not be selected, regardless of how much effort went into them. That asymmetry is real, but it is structural to GSoC.
On the framing that this feels “one-sided”: what we asked of applicants was a 1-2 page preliminary architectural proposal, on top of the standard GSoC materials (CV, motivation letter, availability). That was a deliberately bounded ask, and it is spelled out in the project description and application requirements on our wiki, which we put together precisely so that expectations were explicit upfront. On our side, the mentors put in substantial time across the application window, answering questions on this thread, reviewing the submitted proposals, and running interviews with the shortlisted candidates, on top of committing months of dedicated mentorship and code review for whoever is selected. The exchange is not one-sided; it is that, for any given cycle, only one of many applicants ends up actually working with us. None of this is unique to AiiDA or GSoC; the same dynamic holds for any selective application process.
As general guidance for any future applicant or contributor to AiiDA, or really to any open-source project, two things have come to matter to us more than ever in an era where LLM tools are everywhere: communication and the genuine effort behind a contribution. On the first: the way you have communicated here recently and the messages you left on your closed PR fall short of the professional standard we expect from contributors. We are a small team, reviewing applications and PRs on top of our day jobs, and we are not on call for individual rejection appeals. On the second: fully AI-generated content submitted without real engagement from the contributor actively hurts an application rather than helping it. The sheer volume of AI-generated PRs and architectural proposals we receive around the GSoC deadline makes it hard to keep track of and review everything (and many GSoC mentors have voiced this struggle).
Finally, on the substance of the proposal itself: it was evaluated against the same criteria as everyone else’s, and we are not going to publish a per-applicant scorecard here. We do, however, appreciate the merged PR #7211 as a real contribution.
I am writing this in the spirit you asked for, namely feedback, directly. I hope it is useful for you and everyone else’s future applications.
Best,
Julian