How did we select 30 talks from 280+ talk proposals received in Pycon India 2019?

Detailed Pycon CFP review process from a reviewer

Soumendra kumar sahoo
·Jun 26, 2020·

5 min read

How did we select 30 talks from 280+ talk proposals received in Pycon India 2019?

Subscribe to my newsletter and never miss my upcoming articles

Play this article

I was one of the reviewers for Talk proposals in PyCon (Python Conference) India 2019, which happened in Chennai. This article is written to help someone for submission of CFP in future PyCon India. Especially how their proposals will get reviewed.

There had been around 280 Talk proposals received. We had to limit the numbers to below 30. There was a hefty amount of work involved.

Each selected proposal has to go through three stages.

Stages of a Talk proposal:

  • CFP coordinators review phase

  • Field Experts review phase

  • Final review phase by CFP coordinators and organizers

The narrowing down process:

Almost 150 proposals got rejected due to:

  • No or limited supporting materials.

  • There was not enough content to make to the next stage.

  • No response from the proposer after repetitive reminders.

These activities were carried out by the CFP Organizers. We the reviewers did not get involved on this stage.

The rest of the Talks were categorized into followings:

  1. Game Design and 3D Modelling — Python in developing games, 3-D modelling and animation

  2. Embedded Python and IOT — MicroPython, Python on Hardware, Robotics, Arduino and Raspberry Pi

  3. Culture and society — Diversity, health, productivity, workspace issues, privacy, community building, coding for causes

  4. Others — Everything else that may be of interest to the audience.

  5. Core Python — Language Features, Python Implementations, Extending Python and Standard Library, language internals

  6. Data Science, Machine Learning and AI

  7. Desktop Applications — Qt, GTK+, Tkinter, Gnome, KDE, Accessibility

  8. Scientific Computing — Python usage in scientific computing and research. GIS, Mathematics, Simulations

  9. Developer tools and automation — Testing, CI/CD, Containers, Orchestration, Logging and Monitoring

  10. Web development — Web, Apis, Microservices

  11. Networking and Security — Network Programming, Network Security and Encryption

  12. Each category had their own set of experts, selected by the CFP coordinators based on the applications received to review.

  13. Category Data Science, Machine Learning and AI had received the highest number of proposals, which was more than the sum of next four categories combined. Now do not blame the organizers if you had witnessed more AI/ML talks in the PyCon schedule.

Proposals received across each categories in PyCon 2019Proposals received across each categories in PyCon 2019

  • Review was on FIFO basis, those who submitted earlier, was given feedback earlier and a chance to make their contents worthy to qualify for next stage.

  • There were beginner friendly talks and Intermediate and expert talks as classified by the proposer, based on that I had to think from an audience point of view.

These are few recommendations we have provided to the proposers based on their CFP:

  • Talks with lot of information which need to be squeezed,

  • Talks which need one demo for acceptance,

  • Talks which need better slide design,

  • Talks which need a catchy title

  • Talks which would have been better if added certain topics, etc were provided as feedback to the proposers.

  • Talks better suited as Lightening talk of 5 minutes and not 20 minutes

  • Talks which were niche and will attract only a specific individuals

  • Talks which were pretty obvious and can be easily learnt over internet

  • Talks about promoting their product

Selection process:

  • Each talk need to be reviewed by minimum two reviewers, there were Talks which received more than 10 votes as well. The average of the reviewers’ vote was used to sort out Talks.

  • Based on the number of attendees shown interest for each category, the Talks were selected across the categories. However, if the Talks were not up to the mark, were dropped.

  • Diversity was considered as a added weight for Talks who had, differently-abled speaker and female speaker.

  • Narrowing down from 130 to 50 proposals were easier compared to 50 to 30. There were many feedback, calls, discussions, voting process involved to select the final 30 Talks. These were the most exciting two days of the entire review process.

  • The experienced organizers were able to weed out the overrated talks.

  • Finally 33 Talks had been selected.

My reviewing process:

  • I was the only reviewer and the fortunate one, who got selected to review proposals across six categories.

  • Thereby I had reviewed around 100 proposals.

  • Initially, reviewing each proposal was taking around 40 to 75 minutes to me. As I got myself habituated with the review process, my reviewing speed increased.

  • I got to learn a bit from each proposal; both from the CFPs which got qualified also which had not. I was searching over internet for the terms/concepts I did not know about to have an idea of the Talk.

  • August 2019 was an exciting month. I have learnt a lot of new concepts and gained industry knowledge.


  • There were talks which I would have love to attend at the conference. However, the competition from other proposals were so intense that, these talks could not made it. Heart broken.

  • There were last minute drop-outs from proposers till the penultimate day of the conference. I respect the wait-list proposers who were able to pull this through in a single day.

As a reviewer I have tried to be neutral and unbiased. If you have liked/disliked the Talks in PyCon 2019, now you know whom to blame ;)

Thank you and all the best if you are submitting a proposal.

Further reading

Did you find this article valuable?

Support Soumendra kumar sahoo by becoming a sponsor. Any amount is appreciated!

See recent sponsors Learn more about Hashnode Sponsors
Share this


Views are my own and not represent of my employer.
All articles and images are CC-BY-SA-4.0 licensed unless or until explicitly specified.