Update README.mdMithusa KAJENDRANauthored9ba60468

Learn more.

On This PageAbout UN’s OICTWhy Participate With Us?Key Information for ApplicantsTime commitmentApplication processTips for Getting SelectedChoosing the Best GSoC Project for You2026 Project Ideas & Source CodeFireForm: An Open Source “Report Once, File Everywhere” System for First RespondersExpected OutcomesContributor SkillsEstimated EffortMentorsLicenseOG–CLEWS: Integrating Open-Source Economic and Environmental Models for Sustainable DevelopmentExpected OutcomesContributor SkillsEstimated EffortMentorsLicenseAdditional Contact Information

README.md29.15 KiB

On This Page

About UN's OICTKey Information for ApplicantsChoosing the Best GSoC Project for You- 2026 Project Ideas & Source Code Additional Contact Information

About UN’s OICT

The Open Source Team at the United Nations, based within the UN Office of Information and Communications Technology (UNOICT), supports the development and adoption of open source technologies across the UN system and in collaboration with external partners. The team works to advance digital public goods that address global challenges such as climate resilience, sustainable food systems, and inclusive digital participation. Through initiatives including “Reboot the Earth” and other open source collaborations, our team enables early-stage ideas to mature into scalable, production-ready solutions. Our work emphasizes open collaboration, responsible use of emerging technologies such as AI, focusing on deployment in real-world, resource-constrained environments. The Open Source Team also provides technical guidance, mentorship, and governance support to ensure projects not only align with UN values but also prioritize long-term sustainability.

Disclaimer: All Summer of Code work is conducted independently. Contributors are not considered United Nations employees or official representatives.

Why Participate With Us?

Working with the UN OICT Open Source Team means contributing to open source software with real, global impact. Our projects are digital public goods used by first responders, governments, and policymakers to address challenges such as public safety, climate resilience, and sustainable development.

As a GSoC contributor, you will work on production-oriented systems designed for real-world, resource-constrained environments. Whether improving tools that reduce administrative burden for firefighters or integrating economic and environmental models used in national planning, your work will directly support people and institutions worldwide.

Our projects reflect the shared values of open source and the United Nations: transparency, collaboration, responsible use of technology, and long-term sustainability. You will be mentored by experienced maintainers, gain hands-on experience in distributed open source teams, and contribute code that remains in active use beyond the program.

Key Information for Applicants

Google Summer of Code is best thought of as an “open source apprenticeship”: contributors will be paid by Google to work under the guidance of mentors from an existing open source community. It's a really great opportunity to build new skills, make connections in your community, get experience working with a larger and often distributed team, learn, and, of course, get paid.

Time commitment

Contributors are expected to work either 350 hours (full-time equivalent) or 175 hours (part-time equivalent) over the course of the program. The default schedule runs over 3 months and can potentially be spread over a longer period. You should not plan to take on another full-time internship, job, or schooling during the GSoC period; however, a few weeks of overlap is often fine. Discuss any concerns with the potential mentor for your project of interest.

Application process

To apply, you need to take a look at the mentoring organizations and the ideas that they are willing to sponsor. Typically, you'll choose one of their ideas and work with a mentor to create a project proposal that's good for both you and your chosen open source community. Sometimes, projects are open to new ideas from GSoC contributors, but if you propose something new make especially sure that you work with a mentor to make sure it's a good fit for your community. Unsolicited, undiscussed ideas are less likely to get accepted.

Note that UN-OICT is an "umbrella organization" which means that our team is actually a group of projects that work together to do Google Summer of Code. If you're going to apply with us, you'll need to choose from one of those teams, because that defines which mentors will be helping you with your applications. Contributor applications without any identified sub-org and mentor to evaluate them will be rejected. You can communicate with and even make early contributions to more than one sub-org while you figure out which project you’d like to apply to.

Once you've narrowed it down to a project idea or two, use the application checklist to prepare your project proposal. (Google allows you to submit up to three proposals, but remember you can only be accepted to one organization/project.)

All applications must be sent through the Google system, and must be finalized and fully submitted before the deadline. Don’t wait until the last minute.

Tips for Getting Selected

Google intends Summer of Code to be a way for new contributors to join the world of open source. The contributors most likely to be selected are those who show a likelihood to be highly engaged with the project and continue their involvement for more than just a few months. For most projects, it's more valuable to try to be a good community member than it is to already be a good coder!

Always read all instructions carefully. A large number of applicants don't read the instructions when submitting proposals, and their applications get rejected. For example, every year GSoC organization admins reject a number of contributors who submitted a personal resume/CV, or seemingly random scientific paper, presentation or other file that doesn't contain any information about the project they would like to complete. Sometimes we get dozens of nearly identical form letters from a single university that wind up marked as spam. Don't be one of these! Make sure your application is highly relevant and customized for each project idea application.

Listen and use feedback from others. Every year, we reject a few contributors who simply wouldn't listen to their mentors. Remember: the mentors are using their interactions with you to figure out if it's worth their volunteer time to work with you. No one wants to have an intern who doesn't listen, and contributors who don't listen also don't produce code that the open source project can use, so contributors who don't listen don't get hired. Nor do contributors who are arrogant jerks, or who violate a Code of Conduct for a project or GSoC itself. Be professional, and show that you will take the mentoring relationship seriously.

Here are some resources so you can read up more on how to be an awesome GSoC contributor:

The GSoC Student Guide-- This is a guide written by mentors and former contributors. It covers many questions that most contributors ask us. (Note that it was written when all GSoC contributors were students.) Please read it before asking any questions on the mailing list or IRC if you can! If you are new to GSoC and/or open source, you may specifically find value in reading the“Am I Good Enough?”section. - Google's GSoC Resource List-- Note especially theFrequently Asked Questions(FAQ) which does in fact answer 99% of the questions GSoC mentors and admins have received over the years.

Choosing the Best GSoC Project for You

Choosing a project is a pretty personal choice. Many organizations participate in GSoC each year, not just the United Nations. And each organization may have one or more projects that seem interesting to you. You should choose something you want to work on, and no one on the United Nations team can tell you exactly what that would be! You won’t hurt our feelings if you think a different project or organization is better for you. But, consider a few general questions you might want to ask yourself to help figure that out:

**What software do you already use?**If you use the software, you know a lot more about it and may have ideas about what would make it better! - **What would you like to learn?**GSoC is meant to be a bit of a learning opportunity. Have you always wanted to be more involved with a certain topic? See which projects might help you improve your skills. - **Who do you like working with?**Hang out where the developers do and get to know some of your potential mentors. Which developers inspire you? - **How do you want to change the world?**Do you want to help people learn more? Communicate better? Understand our world better? Lots of python projects can help you do social good! - **How do you like to communicate?**Do you like real-time chat communication like IRC or Matrix? If so, you might want to look for a project with mentors close to your time zone. Do you like asynchronous communication via email or forums? Find a project that works using these tools. Communication is a big part of GSoC (or for that matter, any team-based open source project!) so finding a project that works the way you want to work can make the program more enjoyable for you.

If you're chosen as a GSoC contributor, you're going to be expected to make some decisions on your own, so you can make a better first impression on potential mentors by demonstrating that you're able to do a good job managing the decisions about your GSoC application!

2026 Project Ideas & Source Code

[!tip] Congratulations, you're early!

GSoC 2026 pages will be finalized at the beginning of February 2026, pending our possible acceptance for participation in the program.Applicants should review all information on this page for details about working with OICT’s projects. The projects below are the current candidates for participation in Google Summer of Code 2026. If OICT is selected to participate, the number of contributors we can accept will be based on the allocation provided by Google, which may influence how many projects will participate under the OICT “umbrella” this year. Watch this page for updates in the coming days.

FireForm: An Open Source “Report Once, File Everywhere” System for First Responders

FireForm is an open source, technology-agnostic system designed to reduce administrative overhead for first responders. It is a Digital Public Good (DPG) that enables emergency response agencies; such as fire departments, to eliminate redundant paperwork and reclaim valuable operational time.

FireForm allows responders to report an incident once, using natural language (voice or text), and automatically generate all required agency-specific reports. By leveraging locally deployed, open source AI models, FireForm transforms unstructured incident descriptions into structured data that can populate existing PDF forms used by multiple agencies. The system is designed for deployment in privacy-sensitive, resource-constrained environments and can be adopted or customized by departments worldwide.

First responders are frequently required to document a single incident across multiple agencies, such as fire departments, law enforcement, emergency medical services, and county or state authorities. Each organization maintains its own reporting formats, forms, and data requirements.

As a result, firefighters and other responders often spend hours at the end of a shift repeatedly entering the same information into different systems. This administrative burden reduces time available for rest, training, and emergency readiness, and increases the risk of errors and inconsistencies in reporting.

FireForm provides a centralized “report once, file everywhere” workflow, designed to address the heavy administrative burden faced by firefighters and other first responders. A responder records a single voice memo or completes one master incident description using free-form text. This input is then transcribed, if necessary, and processed by a locally running, open source large language model (via Ollama) to extract structured incident data into a standardized JSON format. The structured data is subsequently mapped to agency-specific PDF templates and automatically fills all required forms. All processing occurs locally, ensuring that no incident data leaves the machine or organization.

This system is particularly important for firefighters because reporting requirements are often time-consuming, repetitive, and prone to error, taking critical time away from emergency response, training, and recovery. By significantly reducing paperwork, FireForm allows responders to focus on saving lives and protecting communities while ensuring accurate, consistent reporting across multiple agencies. The result is not only a reduction in administrative workload but also enhanced operational efficiency, data reliability, and overall safety for both responders and the public they serve.

Expected Outcomes

• By the end of GSoC, the contributor is expected to deliver a stable and well-documented FireForm core pipeline that converts natural language input into structured JSON output, along with a configurable PDF template mapping and auto-fill system capable of supporting multiple agency forms. The project should demonstrate improved AI extraction accuracy through schema validation and robust error handling, ensuring reliability in real-world use.
• In addition, the contributor will provide clear developer and user documentation covering setup, customization, and deployment, as well as unit and integration tests for core functionality. The resulting codebase should be modular, maintainable, and designed to support long-term sustainability and ongoing community contributions.

Contributor Skills

This project requires contributors with strong general programming ability, an interest in open source civic or public-sector software, and the capacity to work across AI integration, automation, and document processing layers. While all contributors will engage with the full system, we expect applicants to demonstrate particular strength in at least one of the following two areas.

All applicants should demonstrate:

• Strong proficiency in Python
• Experience with structured data processing and transformation (e.g., JSON, dictionaries, pandas)
• Familiarity with document processing, PDF generation, or form automation
• Ability to design and implement reliable data pipelines that integrate AI outputs into actionable formats
• Experience writing automation scripts and command-line tools
• Understanding of modular software design and maintainable code
• Experience using Git and collaborative development workflows
• Interest in developing software with real-world impact for public safety and first responders

Applicants with preferred skills (AI/NLP and Data Extraction) should demonstrate:

• Experience working with natural language processing (NLP) or large language models (LLMs)
• Familiarity with AI-based data extraction, transcription, or entity recognition workflows
• Experience designing validation or error-handling for automated data pipelines
• Experience with open source ML/NLP frameworks such as Hugging Face, PyTorch, or TensorFlow

Estimated Effort

Small to Medium (175–350 hours). Final sizing will depend on the contributor’s background and the agreed project scope.

Mentors

Mr. Vincent Harkins(vharkins@ucsc.edu) - Mr. Marc Verges(mvergess@ucsc.edu)

License

This project is licensed under the MIT License.

OG–CLEWS: Integrating Open-Source Economic and Environmental Models for Sustainable Development

• Source Code: https://github.com/EAPD-DRB/MUIOGO

This project aims to integrate two widely used open-source policy modelling frameworks —CLEWS and OG-Core— into a unified, end-to-end decision-support tool for sustainable development planning. By linking sectoral resource systems (climate, land, energy, and water) with a dynamic macroeconomic model, the project will enable policymakers to assess the economy-wide impacts of climate and development policies in a transparent, reproducible, and user-friendly way. These frameworks have been successfully implemented in over 20 countries in a wide range of issues, from informing the development of Nationally Determined Contributions (NDCs) under the Paris Agreement and preventing maladaptation to climate impacts, to assessing options to help lower-income households, or informing on the viability of social protection and pension systems — thus benefitting hundreds of millions of people worldwide.

The Economic Analysis and Policy Division (EAPD) of the United Nations Department of Economic and Social Affairs (DESA) oversees the development and country-level implementation of these open-source modelling tools. For over a decade, these models have supported developing countries facing challenges in achieving sustainable development, particularly Small Island Developing States (SIDS), Land-Locked Countries, and Least Developed Countries. Because they are open source, assumptions are visible and results are reproducible, supporting informed policy discussion. They can be run and adapted at low cost, and they can be calibrated to the data a country actually has, making rigorous analyses feasible in developing and low-income settings where proprietary tools are often out of reach.

The CLEWS (Climate, Land, Energy, Water Systems) framework, built on the OSeMOSYS - Open Source Energy Modelling System, maps the interactions, synergies, and trade-offs among land use, the energy sector, and water systems under climate change scenarios. It allows users to check the actual, physical viability of their plans, from analyzing whether there will have enough land and water to implement a biofuel policy for energy independence, to studying the impacts of carbon taxes on a green energy transition.

The country-adapted OG-CORE is an overlapping-generations (OG) macroeconomic model that enables dynamic general equilibrium analysis of fiscal, demographic, and economic policies over the long term. OG-Core turns government choices into testable scenarios and shows how they shape growth, jobs, and inequality over time and across generations. It can stress-test a range of major reforms, from taxing and spending choices to health or education policies, and technology shifts such as the effects of AI on productivity.

The OG–CLEWS framework expands both models into a single integrated assessment tool by linking their data structures, execution workflows, and analytical outputs. While both models are mature and widely used, they currently operate as separate tools. This project will create a standardized, automated interface between them and a shared execution and visualization layer, enabling integrated analyses that are not currently possible with existing tools. It will provide policymakers with an easy-to-use, accessible tool to check their plans and strategies and back them with an evidence-based, transparent assessment tool. They will gain a more holistic view of their policies, encompassing natural resources management, the energy sector, and their macroeconomic and fiscal implications, in a novel approach that is not currently available to them.

The enhancements developed through this project will be implemented in more than 10 target countries under an ongoing USD 2 million Peace and Development Trust Fund programme, through the through the United Nations Regular Programme for Technical Cooperation, as well as through the Development Account 19th Tranche , ensuring that the results of this work will have real-world impact beyond the GSoC through 2030, helping more countries worldwide achieve their Sustainable Development Goals.

Expected Outcomes

By the end of the GSoC period, contributors will deliver a platform-independent application that supports the standalone execution of the OG-Core and CLEWS/OSeMOSYS models, as well as the integrated OG–CLEWS model. The work centers on extending https://github.com/OSeMOSYS/MUIO (the existing UI for CLEWS) into a unified tool that manages data and scenarios, orchestrates model runs, and presents outputs from both systems.

Cross-platform MUIO: Contributors will refactor and package MUIO so it runs on multiple operating systems beyond Windows, namely MacOS and Linux, preserving all existing functionality. The source code can be found at:

https://github.com/OSeMOSYS/MUIO/releases - OG-Core module in MUIO: Contributors will implement an OG-Core module within MUIO, extending the user interface so that it provides the same features currently available for CLEWS. This includes data input, creating and managing OG scenarios, configuring run settings, launching runs, capturing logs and run metadata, presenting results through interactive graphs, and organizing outputs in a standardized results structure. The user interface will allow users to select whether to run CLEWS or OG within MUIO.

OG–CLEWS coupled module: Contributors will implement a module in MUIO that can execute coupled-model workflows. In coupled mode, MUIO will run one model, apply a provided data exchange pipeline to transform outputs into inputs for the other model, then run the second model. The module will support initially one-way coupling (from CLEWS to OG-Core and vice versa). MUIO will store intermediate exchange files, final outputs and full logs, and will include validation checks that confirm required inputs exist at each step. The user interface will then allow users to select CLEWS only, OG only, or coupled mode. In coupled mode, users will be able to choose linked scenarios, set coupling options (CLEWS to OG or vice versa), run the workflow, monitor progress, and browse outputs for each stage of the coupled execution. The interface will prioritize usability for non-technical users through sensible defaults, guardrails, and informative error messages.

OG–CLEWS converging module: Contributors will expand the coupled module into iterative runs, so that users can launch a simulation that will run the models consecutively (run CLEWS, then feed its outputs into OG-Core and run it, then feed its outputs into CLEWS and run it, etc.) until the simulation converges into a solution (i.e., the results remain approximately the same). The user interface will allow users to select CLEWS only, OG only, coupled mode, or converging mode.

Contributor Skills

This project requires contributors with strong general programming ability, an interest in open-source scientific software, and the capacity to work across model integration, automation, and usability layers. While all contributors will engage with the full system, we expect applicants to demonstrate particular strength in at least one of the following two areas.

All applicants should demonstrate:

• Strong proficiency in Python (3.x)
• Experience with scientific or data-intensive programming (e.g., NumPy, pandas, xArray)
• Familiarity with structured simulation, optimization, or macro/energy models
• Ability to design and implement data exchange pipelines between heterogeneous systems
• Experience writing automation scripts
• Understanding of modular software design and maintainable code
• Experience using Git and collaborative development workflows
• Interest in policy-relevant, sustainability-oriented software

Applicants with preferred skills (Scientific Programming and Model Integration) should demonstrate:

• Experience with numerical modeling, simulation systems, or optimization
• Familiarity with frameworks such as Pyomo, PuLP, GAMS, or similar
• Experience with large, structured datasets (time series, scenarios, sectoral data)
• Ability to design robust input/output schemas
• Experience in building reproducible computational pipelines
• Comfort debugging complex model workflows
• Familiarity with energy systems, macroeconomic, or natural resources models (nice to have)

Applicants with preferred skills (User Interface, Software Engineering, and Packaging) should demonstrate:

• Experience building user interfaces, preferably using Streamlit, Dash, or Gradio, or alternatively React with a Python backend (e.g., FastAPI)
• Data visualization experience (e.g., Plotly, Vega, D3)
• UX awareness for non-technical users
• Experience packaging Python applications
• Experience with automated testing (e.g., pytest)
• Experience with CI/CD pipelines (e.g., GitHub Actions)
• Familiarity with dependency management (pip, conda)
• Familiarity with Docker or containerized workflows (nice to have)

Note on Team Composition: Subject to allocations by Google, we hope to select at least two contributors with complementary strengths: one with a stronger focus on scientific programming and one with a stronger focus on user interface design, usability, and software engineering. However, all contributors must be comfortable working across system boundaries and collaborating on shared design decisions.

Estimated Effort

This project is designed for two GSoC contributors, each committing approximately 175–350 hours. The work is modularized so that contributors can focus on complementary components.

Contributor 1: Scientific Programming and Model Integration Focus Estimated effort: 175–350 hours

Main responsibilities:

• Integrate OG-Core into MUIO
• Design and implement structured data exchange pipelines between CLEWS/OSeMOSYS and OG-Core
• Define standardized input and output schemas
• Implement scenario orchestration logic
• Develop automation scripts to launch coupled and converging model runs
• Build and test reproducible pipelines
• Write documentation for further development and for UI (frontend) integration (i.e., what data outputs are being produced, their content, location, etc.)

Expected outputs (OG-Core MUIO module backend):

• Python-based ETL pipeline between models
• OG–CLEWS MUIO coupled module backend
• OG–CLEWS MUIO converging module backend
• Schema documentation
• Validation tests
• Example country scenarios

Contributor 2: User Interface and Software Engineering Focus Estimated effort: 175–350 hours

Main responsibilities:

• Refactor and package MUIO to work on multiple OS
• Design and implement a policy-maker friendly interface
• Build dashboards for scenario selection, execution, and result exploration
• Implement visualization components
• Add usability features such as defaults, tooltips, and guardrails
• Implement packaging and distribution workflows
• Set up CI/CD pipelines
• Improve documentation and onboarding materials
• Write documentation for backend integration (i.e., what data outputs are expected, their content, location, etc.)

Expected outputs:

• Cross-platform MUIO
• OG-Core MUIO module frontend
• OG–CLEWS MUIO coupled module frontend
• OG–CLEWS MUIO converging module frontend
• Results visualization modules
• Packaged application
• Testing infrastructure
• Deployment instructions

Mentors

Mr. Marcelo LaFleuris a Senior Economist at the Economic Analysis and Policy Division (EAPD) of the United Nations Department of Economic and Social Affairs (UN DESA) who serves as a core maintainer of open-source OG-Core model repositories (e.g., the OG-PHL repository on GitHub) and contributes to public training materials for the OG-Core framework. (lafleurm@un.org) - Mr. Alfonso Acosta Gonçalvesis a Senior Adviser in Integrated Modelling for Sustainable Development at EAPD in UN DESA, lead for the OG–CLEWS modelling project. (acostagoncalves@un.org)

License

This project is licensed under the Apache Licence 2.0 License.

Additional Contact Information

If you have specific questions or want to get started learning more about one of the project ideas above, please contact the mentors listed or use other contact information on the projects’ individual web pages.

If you are stuck and have not received any responses from the mentors listed above, or if you have general questions about participating in GSoC with UN OICT, please contact or OICT Organizational Admins, Mithusa Kajendran (mithusa.kajendran@un.org) or Michael Downey (mdowney@humanitech.group).

Credits: Some content on this page was adapted from "Python Summer of Code" by Python Software Foundation, licensed under CC BY 4.0.