Published February 20, 2026
This year I was asked to join the Program Committee for PyCon US and I want to talk a bit about what I did and how the talks were chosen.
After the CFP closes all of the talks go to the Program Committee (the Charlas, Posters, and Tutorials all go to the various groups that handles them, I only dealt with talks so that's what I'm going to talk about here.) This year there were 741 talks submitted for a total of 76 spots, so just over 1 in 10 of the talks that were submitted were selected. The process of reviewing the talks is done in 3 phases:
The goal in the first phase is to whittle the number down to something that we can reasonably handle. To do this, reviewers answer some questions about the proposals. At this point, the reviewers are seeing 2 things:
The reviewers do not know who wrote the talk. They do not see the "additional notes". Work is done to anonymize the proposals and redact anything that could let the reviewer figure out who wrote the proposal. As we read through each proposal, we answered the following questions:
While there were a lot of amazing proposals that checked most, if not all, of the boxes - siginificantly more than the 76 we could pick for the schedule - there were a lot that fell short. Since ~50% of the proposals don't get past this round, I want to talk a bit about some of the things I saw that didn't make it through. I'm going to go through each of the questions and try to give some insight that can help you (and me!) write a successful proposal for next year.
Thankfully none of the proposals caused me (or any of the other reviewers) to answer "no" to this one.
When answering this question I was primarily looking at the description. Is it understandable. Does it make sense. Are there mistakes in the spelling or the grammar. Do I understand what the proposal is talking about? This one is tricky. We saw a not small number of proposals that seemed to have been written entirely by LLM with little to no human review. The proposals that gave off that vibe tended to get a "no" on this question. Other than that, there were some that didn't make the cut on this question because they were too scattered or just poorly written. I will say, proofreading matters.
For this I looked at the outline. I looked at the timings per section and the topics covered. I tried to get a feel for whether enough time was being devoted to any given section in the outline and if all of the ideas covered in the description actually made it into the outline. Were they requesting the appropriate time for the talk they wanted to give (some asked for a 30 minute slot but the timings added up to more than that, while others asked for a 45 minute slot and the timings would fit into the 30 minute slot.) Some of the issues that I saw (and that other reviewers noted) included:
Now that we've read and evaluated the description and the outline, does it all make sense? Is it discussing a single topic? Is it all over the place? Does the outline match the description?
Since PyCon US is a Python conference we want, as a general rule, talks about the Python ecosystem. But that is something that is very broad. It could be a talk about the Python community, or the open source community. It does not necessarily need to be a technical talk. It just needs to touch Python in a not small way. Some of the ones that failed to hit this criteria were ones that were very general they mentioned Python in their descriptions and outlines but you could replace that with any other language and it would make sense. Those are not the talks I was looking for.
This one, in my opinion, is the big one. Are there enough people that attend PyCon US that would be in the room for it not feel empty. We don't want a speaker talking to just the video camera - we want them talking to people. We want people engaging the speakers and other attendees about the things that they have heard. If we end up with empty rooms for talks that means we probably missed out on selecting other talks that might have filled those rooms.
Now that I've gone through the questions, here's my advice for writing your proposal. Have other people you know in the community, preferably people who are planning on attending PyCon US, read it. Have them proofread it. Show them these questions and get their opinions on how they would answer the questions. When you have consensus that you can get a "yes" on all 6 you're good to go.
I will add one other nugget here, the reviewers have the option to select a talk to pass to phase 2 in spite of the proposal not meeting all of the objective criteria. This year there were a handful of talks that the proposal looked so good to me that I pressed the button on them. I know that some of the other reviewers did as well.
In phase 2 the talks that passed the first phase are grouped into batches based on their topics. Reviewers are added to the batches and review the proposals in each batch noiminating up to 2 for inclusion in the final schedule. The reviewers also have the ability to leave comments on the batch advocating for specific proposals. For me, this phase was reading through all the proposals - now deanonymized and with additional information such as the notes added by the submitter - and try to answer question 6 again, will there be enough audience for this talk. While my general outlook was asking myself, "Would I want to attend this talk?" that wasn't the only thing I looked at, since some of the topics aren't of interest to me specifically.
At this point all of the remaining talks have been scored (the number of nominations being the effective score) and a schedule is built. The reviewers have another chance to look at the schedule as a whole and try and determine if it looks well balanced and to potentially advocate for any talks that might not have made the cut. The hard part here is that there is a list of very good talks so adding one means taking out a different one.
At the end of phase 3 emails are sent out to everyone that submitted a talk telling them what the status of their proposal is.
Personally, I learned a lot throughout this process. I learned how to make my proposals better by getting a better understanding of what the people who review them are looking for. I will say that, according to the committee chair, this year had less discussion than prior years but, overall, I think there is a great selection of talks and there is a lot to look forward to.
And if you made it this far, I highly recommend you read Maggie Fero's thread on treating CFP submissions as offers. It's a great way to frame this whole process in your mind. I can tell you that included in the 675 talks that didn't make the cut there are some gems that I hope get submitted next year or to other conferences because they should be heard.