This page contains general information on European Summer of Code 2025:
- European Summer of Code: background and org structure
- For applicants: how to apply, list of projects 2025
- For projects: how to join, timelines
ESoC webpage - ESoC on LinkedIn
- 2025-04-29: Batch 3 projects released
- 2025-04-28: New Batch 3 projects release date
- 2025-04-10: Batch 2 projects released!
- 2025-03-27: Applicant Guide, Sponsor Guide published
- 2025-03-27: Early Bird Batch of Projects released!
- 2025-05-16: Start of rolling basis projects
- 2025-05-22: New Project release!
- Information on European Summer of Code
- Applicant Guide
- Sponsor and Project Guide
- 2025 Project List - Applied Projects
- 2025 Project List - Open Source Projects
European Summer of Code (ESoC) funds stipends for contributors new to open source, and matches open source projects and applied AI projects throughout Europe. We are much indebted to the Google Summer of Code for inspiration.
The goals of ESoC are threefold:
- provide opportunities for junior developers to engage with open source worldwide
- link the worldwide open source landscape with private and public sector project throughout Europe
- provide support to the wider open source ecosystem
In particular, we expressly help applied projects to connect with relevant open source, open source projects to connect with applied projects and sponsors, and applicants to work on open source directly in an applied context.
- participants apply for projects with open source libraries or applied partners
- public and private sector sponsors submit projects
- open source projects apply for support
- hubs throughout Europe carry out the matching process
ESoC is organized in a decentral and distributed fashion, by hubs. Current hubs are:
- France: probabl - the scikit-learn company
- Germany: German Center for Open Source AI
- Italy: Fondazione Bruno Kessler
We are actively looking to expand our network of hubs.
Get in touch on LinkedIn to discuss details.
We are actively looking to expand our network of affiliated open source projects.
Projects may affiliate with one of our hubs for streamlined operational integration and funding opportunities.
Fill out our open source project onboarding form (link tba) to join!
(While form unavailable: get in touch on LinkedIn.)
Currently affiliated projects:
pixipgmpypykalmanpytorch-forecastingrattlerscikit-basescikit-learnskoreskproskrubsktime
dsip-tsGradient-Free-Optimizershyperactiveprophetverse
- 2025 projects are released on a rolling basis.
- projects have their own timelines and may come with their own terms of conditions
- project specifics are listed in the project card, in the list of released projects below
Applicants need to do two things:
- register for ESoC through the ESoC applicant registration form
- follow any additional "how to apply" steps from the project card of your preferred projects
- The project card may require you to fill out a second form, or complete an additional task
- The project card may also simply say that filling out the ESoC form suffices (as long as you mention the project)
- While there is no limit in the number of projects you can apply for (or penalty), we recommend to focus on at most 3. Applying to too many projects will reduce your time per application and impact application quality.
The deadline for completing both ESoC application form, and the project specific application steps, by the project application deadline. The project application deadline varies by project, and is specified in the project card, below.
- Please ensure your submitted e-mail address is correct. Letters will be sent to this e-mail address.
- Please use the same name and e-mail for any fields and forms throughout the process
- Ensure to note down relevant date periods in advance, and ensure to check your e-mail in those periods.
- In case of submitted forms, most forms allow you to update the submission up until the deadline.
- If this is not the case, simply submit another application under the same name. The last submission received before the deadline for a given project counts.
- note that the last received forms count here - not submitted forms which are not received due to technical error, or received too late.
- So ensure you submit something early, as some submission systems may overload or not work properly in the last few hours before deadlines.
Applicant Q&A in the discussion forum
You will receive a letter from ESoC on your application(s) no later than 6 weeks after the application deadline of the project you applied to.
This can be:
- acceptance to one or multiple projects. This means you are offered a slot.
- waitlist status for one or multiple projects. This means you may be offered the slot if higher ranked applicants withdraw.
- In this case a follow-up letter with acceptance or rejection will be sent at a later time.
- rejection from one or multiple projects
The letter will be sent to the e-mail address you registered with.
In case of acceptance, you will need to reply to the letter within 1 week of receipt.
Please note: once you accept the offer from a project, any pending applications of yours to other projects in 2025 are automatically withdrawn, including those from later batches where the outcome may not yet be clear, or projects where you are on the waitlist.
- Private and public sector organisations wishing to host a project should submit a project sponsor form. One of our hubs will be in touch shortly after submission.
- open source projects should directly contact our LinkedIn organization (to be replaced by a form)
One of our hubs will work with you to prepare your project for release.
Finalization will require:
-
choosing a funding model:
- option A: donation for stipend. This is easiest, with low contractual complexity: you donate to a hub, and the hub contracts with the participant. The "stipend" requires that a significant portion (but not all) of the outputs are in the open domain, e.g., contributions to an open source project or a research report.
- option B: paid internship. You can use your own contract template, in this case the hub will only act as matchmaker, and you will diractly contract with applicants and manage any tax and visa requirements.
-
completing the project card, see details below and example project cards
- this includes finalizing an application process with your hub. "Default" processes are available, these select for generic AI, python, and open source capabilities.
- for data driven projects, it will also require finalizing data batches and data sharing processes before the project starts
Project cards should include:
- project title, project logos
- one-sentence description
- paragraph: project goals
- paragraph: about the organization (optional: logo)
- paragraph: technical details (very short - link details)
- clear "HOW TO APPLY" info, deadlines
- optional: eye catcher image
- optional links to: more details, data (where applicable), videos or presentations
Some examples are in the project list below.
Build the open source AI landscape for causal AI in time series, with applications across supply chain management and retail!
Project overview (link to detail info sheet)
Causal time series models are central to many industry grade use cases in production, retail, and supply chain.
Join Electrolux and the German Center for Open Source AI to start building an interoperable, open-source ecosystem of causal tools for forecasting and structure identification - in collaboration with pgmpy and sktime.
Electrolux Group is a leading global appliance company. Products include major appliances and vacuum cleaners intended for home consumer use.
Within Electrolux Group, the AI portfolio includes use cases like prescriptive marketing, production optimization, and dynamic pricing.
HOW TO APPLY: follow the GC.OS applied project application guide. Also make at least one substantial pull request to pgmpy, sktime, or - optimally but not necessarily - both.
- Deadline: June 12, 18:00 UTC
- Hub: German Center for Open Source AI
- Sponsor: Electrolux
- T&C: pro-rated stipend disbursed by German Center of Open Source AI.
Previous projects are listed below for reference.
Please do not apply - these stipends no longer accept applications.
For open source projects:
- you are, of course, invited to contribute to any open source project at any time!
- other opportunities may be available with the respective open source project directly, e.g., mentoring.
APPLICATIONS CLOSED
Build open source AI for drug discovery, accelerating aptamer design for diagnostic and medical applications! Sponsored by German engineering services provider ecoSPECS.
Project overview (link to detail info sheet)
Aptamers are highly specific biomolecules, which can pave the way for individualised diagnostics and therapy, allowing personalized treatment of various diseases like cancer, infections or autoimmune disorder. Classically done through in-vitro (lab) experiments, AI-driven in-silico approaches have seen promising advances, fuelled by the success of AlphaFold.
Join ecoSPECS and the German Center for Open Source AI to start building an open source ecosystem with reusable AI software tooling, interoperable standards, and data benchmarks!
ecoSPECS is a German engineering service provider, specializing in services such as qualification, validation, consulting, training, and project planning across various industries, with a focus on pharmaceuticals, engineering, and food production.
HOW TO APPLY: follow the GC.OS applied project application guide
- Deadline: May 14, 18:00 UTC
- Hub: German Center for Open Source AI
- Sponsor: ecoSPECS
- T&C: pro-rated stipend disbursed by German Center of Open Source AI.
APPLICATIONS CLOSED
Build AI for design and equipment qualification of cleanrooms for biotech and pharmacological applications! Work with German engineering services provider ecoSPECS to fundamentally accelerate qualification processes in drug development, production, and monitoring worldwide.
Project overview (link to detail info sheet)
Cleanroom design for biotech applications involves numerous "boilerplate steps". For instance, creating documentation in pre-specified format, particularly for identifying design and quality requirements within user specifications.
Although the design process is expert-driven, significant amount of formulaic editorial work is required. AI could allow experts to focus on complex design tasks, rather than spending time on boilerplate. This would lead to significant acceleration and cost savings in the availability of biotechnological and pharmaceutical processes, systems, and equipment.
ecoSPECS is a German engineering service provider, specializing in services such as qualification, validation, consulting, training, and project planning across various industries, with a focus on pharmaceuticals, engineering, and food production.
HOW TO APPLY: follow the GC.OS applied project application guide
- Deadline: April 9, 18:00 UTC
- Hub: German Center for Open Source AI
- Sponsor: ecoSPECS
- T&C: pro-rated stipend disbursed by German Center of Open Source AI. Data sharing agreement required.
APPLICATIONS CLOSED
The framework for causal inference, causal structure learning, and causal simulation.
Contribute to pgmpy - the causal learning package! Help implement new types of causal models, and increase interoperability of, and within, the package!
pgmpy is an open-source project backed by an active community of researchers and developers.
It is the major causal modelling package not affiliated with a single commercial entity, with a distributed user and maintainer base.
GitHub repo · Discord · Documentation
Multiple projects are available: ESoC projects 2025, mentored projects 2025 - you can also suggest your own project, by opening an issue.
HOW TO APPLY: mention pgmpy as an interest on your ESoC application (you can update this if you already applied).
Then, make a pull request on a good first issue by May 14. You will be contacted with further details by the pgmpy team.
- Deadline: May 14, 18:00 UTC
- Hub & Sponsor: German Center for Open Source AI
- T&C: pro-rated stipend disbursed by German Center for Open Source AI
APPLICATIONS CLOSED
Elevate ML Development with Built-in Recommended Practices
Contribute to skore, the scikit-learn sidekick! skore is the Python library designed to enhance ML development by diagnosing methodological errors, while also providing automated evaluation tools that generate reports with plots, and metrics.
Help us improve data visualization in skore by implementing more displays, a Plotly display backend and providing access to underlying display data.
Probabl delivers universal technology enabling data scientists and teams to make the most of their data, leveraging the best practice in machine learning and artificial intelligence.
Probabl is a spinoff from the Inria research lab and co-founded by experienced tech entrepreneurs and core-developers of scikit-learn. Probabl's core mission is to develop and maintain commons for data science.
GitHub repo · Discord · Documentation
HOW TO APPLY: mention sklearn or skore as an interest on your ESoC application (you can update this if you already applied).
Then, make a pull request on a good first issue by April 26. You will be contacted with further details by the skore team.
- Deadline: April 26, 18:00 UTC
- Hub & Sponsor: probabl
- T&C: internship with probabl
APPLICATIONS CLOSED
Package Management Made Easy
pixi is a cross-platform, multi-language package manager and workflow tool built on the foundation of the conda ecosystem.
Projects are available to improve user experience, build user interfaces, and extend scope to more programming languages!
pixi is maintained by an open developer community and the Berlin company prefix.dev.
Its mission is to build a simple package manager for all programming languages.
See here for a list of projects ideas.
HOW TO APPLY: mention pixi on your ESoC application (you can update this if you already applied).
Then, make a pull request on a good first issue by April 26. You will be contacted with further details by the pixi team.
- Deadline: April 26, 18:00 UTC
- Hub: German Center for Open Source AI
- Sponsor: prefix.dev
- T&C: pro-rated stipend disbursed by German Center for Open Source AI
APPLICATIONS CLOSED
Rust crates for fast handling of conda packages
rattler is a library that provides common functionality used within the conda ecosystem.
Projects are available to implement build process mechanisms, for instance via parallelization
The goal of rattler is to enable programs and other libraries to easily interact with the conda ecosystem without being dependent on Python. Its primary use case is as a library that you can use to provide conda related workflows in your own tools.
See here for a list of projects ideas.
HOW TO APPLY: mention rattler on your ESoC application (you can update this if you already applied).
Then, make a pull request on a good first issue by April 26. You will be contacted with further details by the rattler team.
- Deadline: April 26, 18:00 UTC
- Hub: German Center for Open Source AI
- Sponsor: prefix.dev
- T&C: pro-rated stipend disbursed by German Center for Open Source AI
APPLICATIONS CLOSED
Simple and reliable optimization with local, global, population-based and sequential techniques
Gradient-Free-Optimizers provides a collection of easy to use optimization techniques, whose objective function only requires an arbitrary score that gets maximized. Gradient-Free-Optimizers is the optimization backend of the optimizer toolbox hyperactive, but it can also be used by itself as a leaner and simpler optimization toolkit.
GitHub repo · Discord (shared with sktime)
See here for a list of projects available.
HOW TO APPLY: fill out the sktime application form and submit a project proposal for hyperactive and/or GFO.
- Deadline: April 26, 18:00 UTC
- Hub & Sponsor: German Center for Open Source AI
- T&C: pro-rated stipend disbursed by German Center for Open Source AI
APPLICATIONS CLOSED
The unified framework for machine learning with time series
Contribute to sktime! sktime offers a unified API for forecasting, classification, detection, and other learning tasks.
Exciting projects are available in the areas of deep learning, foundation models, probabilistic forecasting, and more.
sktime is an open-source project backed by an active community of researchers and developers. It is openly governed, and follows a mission of technical integration and content neutrality.
GitHub repo · Discord · Documentation
Multiple projects are available: project ideas 2025
HOW TO APPLY: follow the sktime application guide
- Deadline: April 7, 18:00 UTC
- Hub & Sponsor: German Center for Open Source AI
- T&C: pro-rated stipend disbursed by German Center for Open Source AI
APPLICATIONS CLOSED
Deep learning for time series forecasting made easy with PyTorch
pytorch-forecasting is a Python library built on PyTorch, designed to simplify training and using deep learning models for time series forecasting.
Projects are available, focusing on improving model interfaces, upgrading the package to include foundation models, and more!
pytorch-forecasting is maintained by the sktime community in collaboration with GC.OS and FBK (Fondazione Bruno Kessler, ESoC hubs), a merge with dsipt-ts is planned for the upcoming 2.0 release.
In the dedicated card you can find more details about the project, some of the activities are taken and grouped from project ideas 2025.
HOW TO APPLY: follow the sktime application guide - this includes projects for pytorch-forecasting.
- Deadline: April 7, 18:00 UTC
- Hub & Sponsor: Fondazione Bruno Kessler
- T&C: pro-rated stipend disbursed by German Center for Open Source AI