PAPIs 2018 Oct 16 - 17, 2018
CFP closed

The CFP closed on Sep 16, 2018 at 11:59pm GMT

Thank you for all submitted proposals!

CFP Stats

50 proposals

PAPIs is the international series of community conferences dedicated to real-world Machine Learning applications. Previous editions took place in London, São Paulo, Sydney, Barcelona, Paris, Valencia, and this year's global edition (PAPIs 2018) will be held in Boston for the 3rd consecutive year. Engineers, managers, scientists, hackers and researchers will meet to share experiences and discover the latest innovations, techniques and tools to build ML applications.

Key dates

  • July 5 (23:59 GMT): Main call for proposals closes
  • July 16: Notifications of acceptance/rejection due
  • July 23: Author confirmations due
  • September 16: Lightning Talk and AI Startup Battle application deadline
  • October 16-17: Conference takes place in Boston
  • November 16: Full papers due for proceedings (optional)

We'd like to strongly encourage women to submit proposals. As a community conference, we think it's important to increase diversity in AI. Applications for funding to reimburse childcare and travel costs will be invited from women and primary carers whose proposals are accepted. We're also looking for sponsors to support our diversity efforts!

Sign up here to get updates!


ML Engineering & ScienceOps (technical) — October 16

This track is targeted at Software Developers, ML Engineers, and ScienceOps professionals (DevOps for Data Science). It is dedicated to discussions on the technical challenges of deploying ML systems in production, and lessons learnt from experience. We invite industry participants to describe their war stories and the architectures, pipelines, frameworks, APIs, etc., that they built to create or improve intelligent applications. Proposals for this track should be technical. They can range from ML beginner-friendly how-tos to deep dives for more experienced professionals. Presentations will be 20' long, plus 5' for Q&A.

Applying ML in the real world — October 16

This track is for everyone who's interested in the integration of ML into real-world applications — from engineers and scientists to managers. It will feature presentations of ML use cases in new areas, discussions on the challenges of integrating ML into various applications/products, and lessons learnt. This can include domain-specific challenges, as well as those related to product design, transparency, privacy, fairness, accountability, ethics, etc. We would like to also encourage the community to share their failure stories — after all, you aren’t a real ML innovator until you’ve had to deal with massive failure of your ML system! Presentations will be 20' long, plus 5' for Q&A.

New ML techniques (technical) — October 17

This is our second technical track, targeted at the most experienced practitioners who have an advanced knowledge of ML. It is dedicated to the presentation of innovations, applied research, and new ML techniques that have the potential to take the industry further. Proposals for this track will be evaluated on the basis of their novelty. Presentations will be 20' long, plus 5' for Q&A.

Tools — October 17

This track will feature 10' demos and 30' tutorials of tools and APIs that make it easier to experiment with ML, integrate it into real-world products/applications, or to deploy it at scale. If you need to show slides during your presentation, please limit them to the strict minimum. We particularly welcome tools which are open source, or freely accessible. We will also select demos of proprietary tools as long as the presentations are not commercial, they are engaging, practical, or teach how to do a specific task with the tool being demonstrated.

What we're looking for


We're looking for practical and focused presentations, on a wide range of experience and technicality levels. We’re looking for a diverse and creative line-up of speakers, preferably with experience in public speaking. Proposals will be blind-reviewed by our Program Committee and evaluated on the basis of their significance and/or clarity of presentation. Please consult our tips & rules for more detailed information and advice on how to craft a good proposal and submit it.

Presentations should be non-commercial. A very limited number of sponsored speaking slots are also available — please contact the sponsorship team for more information.


We welcome technical papers, industrial experience reports, review papers and research papers. While submitting a paper is no requirement to be accepted to present at PAPIs, we highly encourage authors to take the time to share their work with the community in this format as well. Papers presented at PAPIs are published in Proceedings of Machine Learning Research (see our 2016 and 2015 proceedings as examples). More information is available in our tips & rules.


PAPIs hosts the world's 1st startup competition where participants are judged by an AI on stage — not a human jury! We're looking for new startups that have AI / ML at their core. Selected startups will be able to present on stage and exhibit at the conference. Applications should be made via a separate platform (To Be Announced).

We’re looking forward to receiving your best proposals!

If you have questions or if you have not heard about the status of your submission, please contact us.

Program Committee



This website is based on rubycentral’s cfp-app. These guidelines were inspired from those of our previous conferences — many thanks to everyone who provided feedback and to all of our previous Program Chairs: Francisco Martin, Florian Douetteau, Keiran Thompson, Alastair Moore, Paula Fadul, Thierry Silbermann.