Find open source projects actively accepting contributors. Search repositories, filter by program milestones, difficulty tags, or tech stack.
Use our Orbit AI Matcher to find out! Get instant matching scores based on your developer skills, preferred frameworks, and contribution experience.
Convert your selected open-source project into a winning GSoC, LFX, or Outreachy application using Proposal Studio.
The objective of the project is to develop a Chrome extension with the purpose of identifying fraudulent information, thereby enhancing online safety and fostering greater productivity for internet users. The project entails a series of stages, primarily focused on the utilization of Natural Language Processing methodologies. During this season, a context detection functionality will be integrated, leveraging the capabilities of the GPT-3 API to provide users with comprehensive insights into news articles, empowering them to make well-informed decisions. The final output will be the updated web service and extension, which will be ready for publishing. The project also aims to enhance the SSL validator and Security Header to ensure maximum security. To achieve this, the codebase will be migrated from Manifest Version 2 to Manifest Version 3, and the model must be integrated into the Chrome extension. The final product will provide users with a reliable tool to identify fake news and ensure that they receive accurate information while browsing the internet.
The aim of this project is to add support for an LLM (large language model) based on tools such as GPT-4 to support AI-based programming support for our users. This support could range from proposing code from text prompts, help in correcting bugs or other issues, finding a good architecture or meaningful names, writing Catrobat language tests for one's code, proposing extensions, proposing project ideas, explaining downloaded code, providing translations of variable or message names etc from downloaded projects of other users, proposing image and sound resources based on the context and text prompts, explaining basic or more advanced concepts needed to realize projects, e.g., simple concepts such as variables and loops, or more complex ones such as software design patterns or testing strategies. Overall, the proposed project aims to provide a more comprehensive and tailored solution for Catrobat users, with a broader range of features and a more user-friendly interface compared to existing products.
<p>I want to extend the tutorial offered by LabLua for the Ceu language. As it is, when you write code on their website and press submit, the code is sent to their server, compiled and the result is sent back. The problem with this, is that it doesn't show the entire capabilities of the language. For instance if you write a program that prints a message every second for 10 seconds, it will return that message printed ten times in an instant. What I want to do by using Emscripten is to have a real time tutorial, so that after the server sends back the result, it will actually print the message every one second, for ten seconds. I will also want to create a second tutorial that implements SDL as well. SDL is a language that can create shapes and interacts with keyboard inputs and the mouse. There already is compatibility between Ceu and SDL and there is also compatibility between SDL and Emscripten. The challenge is to create a synergy with all three languages. I believe this to be feasible, because you can transform a Ceu program into a C program that can then be compiled with Emscripten.</p>
The current OWASP Snapshot system is passive and requires manual updates, forcing community members to constantly check the website to see what has changed across the organization. This project solves that problem by transforming snapshots into an active, automated digest platform. I will build a system that automatically collects weekly updates expanding the scope to include events, blog posts, and pull requests and delivers them directly to interested users. Contributors will be able to customize their subscriptions, choosing exactly which projects and data types they want to follow. I will also redesign the website's snapshot pages to support date-range filtering, full content display, and social sharing. The core architecture will prioritize GraphQL for all frontend needs. The primary deliverables are the automated processing pipeline, the personalized email newsletter system, subscription management, and the enhanced web interface. A snapshot feedback system and supplementary REST endpoints are planned as stretch goals.
<p>Elements of differential geometry enable us to quantify the curvature of a surface. The core elements of biological membranes, phospholipids, provide tridimensional configurations from which a surface can be derived to calculate curvature descriptors. We would like to integrate to MDAnalysis an analysis module to calculate averaged mean and Gaussian curvature from Molecular Dynamics simulations. By integrating a membrane curvature analysis module in MDAnalysis, users will benefit from a tool that enables rapid extraction of relevant properties of lipid bilayers in biomolecular systems. Our approach extracts key elements of the membrane using phospholipid head groups to define a surface, followed by a transformation into NumPy arrays. In this way, the functionality offered by NumPy and SciPy can be used to derive values of mean and gaussian curvature of biological membranes. Since MDAnalysis already works very well to explore data interactively, visualization of the membrane curvature analysis outputs in 2D-maps can be easily performed.</p>
While Jenkins is the backbone of modern CI/CD, troubleshooting failed builds and complex configurations remains a time-consuming bottleneck. This project develops a Diagnostic AI Chatbot plugin to drastically reduce this friction, minimizing debugging time and maximizing developer productivity. Powered by an advanced Retrieval-Augmented Generation (RAG) architecture, the agent leverages LangGraph and hybrid search to intelligently filter noisy build logs, cross-referencing them with official documentation and community discussions in order to deliver precise root-cause analysis and actionable fixes. Architecturally, it utilizes a decoupled FastAPI backend, ensuring zero computational overhead on the Jenkins Controller. The system is highly flexible: it is designed with a privacy-first approach optimized for local open-source LLMs, while seamlessly supporting integration with third-party commercial APIs. This allows administrators to effortlessly toggle between absolute data privacy and frontier model performance based on their infrastructure needs.
<p>In the last GSoC, kingiw created an initial version of casbin.js. This version has a CASL-like API. He also adds support to go-casbin and node-casbin.</p> <p>However, casbin.js relies on an insufficient way to share permissions between the frontend and backend. The backend needs to send the whole policy definition model to the frontend, which will use node-casbin to check the permission. This method requires a heavy data transfer in between, and has potential security problems.</p> <p>To build a production-ready frontend permission control library, we need to</p> <ul> <li>Design a new data structure to only store one-user permission.</li> <li>Add a full set of API to operate on the data structure.</li> <li>Implement integration with the frontend frameworks like React or Vue.js.</li> <li>Add util function in different implementations of Casbin.</li> </ul>
<p>The Beacon specification allows to query whether a genomic variation is present in a repository. But the current implementation is based on an outdated version of this specification. Many new parameters that are useful for filtering down the data which can improve the querying experience are available according to the latest specification. To improve this querying experience and search results, I would be updating the existing implementation, so that it matches with the current specification.</p> <p>Also, current variation search APIs retrieves all the fields together. This is a lot of information at once and is cluttered. So, I would be redesigning these APIs and would make it closer to RESTful principles.</p> <p>Now, variation search by 'geneId' is not efficient. To deal with this issue, I would build a data pipeline that would load mappings from a 'geneId' to its coordinates. These coordinates would then be used to search for variations.</p> <p>I would be also be adopting the models defined by VMC which would enable standardized exchange of genomic data.</p> <p>Based on the work done, feed backs will be provided to GA4GH workstreams.</p>
<p>The VideoLAN organisation has transformed the digital audio and media space since its inception over 10 years ago. The ease of use of its forerunning product, VLC Media Player, has made it the number one choice for millions of users around the world for viewing and interacting with media. Yet, despite this VLC Media Player’s design is relatively outdated. Features should be readily obvious to users, not hidden in multiple menus so that they can be accessed quickly and easily.</p> <p>As an avid VLC Media Player fan, this forms the motivation behind my proposal. I propose to update VLC Media Player by leveraging Qt/QML to design a more modern and elegant interface. Specifically, I plan to redesign and implement a new welcome screen which is visible upon first opening of VLC Media Player. This includes the relevant Privacy and Metadata Consent passages for users as well as the ability to alter some commonly changed settings through an easily accessible modal.</p>
InVesalius is an open-source software that enables synchronous 2D and 3D visualization of medical images, facilitating the reconstruction of computed tomography (CT) and magnetic resonance imaging (MRI) data into three-dimensional models. It allows users to create masks easily through various preset parameters, such as predefined threshold values for different tissues and materials. However, in cases where automatic segmentation is insufficient, manual editing becomes necessary. This process can be quite slow and labor-intensive, as it requires editing each slice individually, which is particularly time-consuming for large volumes. Implementing a feature that enables mask creation through projections directly on the 3D rendering screen would significantly enhance InVesalius's usability and efficiency. Currently, mask editing in InVesalius requires a slice-by-slice approach, which can be repetitive and prone to errors. By allowing users to edit the entire volume simultaneously via 3D projections, the segmentation process would be streamlined, saving time and increasing accuracy.
BLT University is an interactive security learning platform built on top of OWASP BLT that transforms existing labs into hands-on, code-driven exercises. It migrates the current Django-based labs into a static Jekyll system where content is stored and rendered as Markdown, making it more scalable and easier to maintain. Instead of static, theory-heavy content, users work directly with vulnerable code through a structured Identify → Explain → Fix workflow to understand how vulnerabilities appear and how to fix them in practice. The project also includes GitHub Actions workflows for validation and syncing, along with a security intelligence pipeline that analyzes GitHub issues, pull requests, and review comments to identify recurring vulnerability patterns. These insights power a dashboard and a recommendation system, helping connect real-world security issues with relevant labs and improve hands-on learning for contributors. It aims to help contributors build a practical security mindset while improving the quality of open-source contributions.
<p>CMS is a general-purpose detector at LHC. During a run, it generates about 40 TB data per second. Since It is not feasible to readout and store such a vast amount of data, so it selects and stores only interesting events or events likely to reveal new physics phenomena. Currently, Boosted Decision Trees (BDTs) with handcrafted features are used for that.</p> <p>The goal of the project is to test the performance of different deep learning algorithms and minimize the latency in inference. The best models in each algorithmic type will be wrapped into a functional prototype emulator. Other than this, a significant emphasis will be given on proper documentation.</p> <p>Links - <a href="https://github.com/PRATEEKKUMARAGNIHOTRI/Muon-Momentum-Estimation-and-Particle-Identification" target="_blank">Submitted pre-tasks</a> | <a href="https://drive.google.com/file/d/1mzLp3aWz1Kb2UADOphoS1ILmlzo35BPI" target="_blank">CV</a> | <a href="https://github.com/PRATEEKKUMARAGNIHOTRI/" target="_blank">GitHub handle</a></p>
Specialization is a compiler transformation that optimizes generic program code such that code using primitive types can execute without significant overhead. The @specialized annotation in the Scala 2 programming language allowed programmers to direct the compiler to create optimized generic data structures for each primitive type. @specialized is used significantly in the implementation of high-performance Scala 2 libraries, allowing for performance without the productivity of generics. This optimization trades performance at the cost of extra space, as generic code must be specialized for each of the primitive types, scaling exponentially on the number of type parameters. Since the Scala 3 programming language has evolved significantly from Scala 2, the @specialized annotation is no longer available. This proposal will prototype a new scheme of specialization for the Scala 3 compiler based on inline traits that avoids the performance pitfalls of the Scala 2 implementation, enabling the evolution of many high-performance Scala 2 libraries into Scala 3.
The Unicode Consortium maintains extensive locale datasets critical for internationalization tasks such as date/time formatting, text segmentation, measurement conversions, and collation. Manual verification of these datasets is challenging due to their vast scale. My project aims to address this by leveraging large language models (LLMs) to automate and enhance the validation of locale data. The approach involves designing targeted queries for an LLM, systematically comparing the generated outputs with existing CLDR/ICU datasets, and identifying discrepancies. These results will then be processed to isolate significant issues, and automated tools will be developed to flag potentially erroneous data for translators. Deliverables: 1.A robust pipeline for querying LLMs and generating locale data predictions. 2.Comprehensive comparative analysis between LLM predictions and CLDR/ICU datasets. 3.An automated issue detection and reporting mechanism integrated with existing Unicode systems. 4.Thorough documentation and guidelines for future dataset validation workflows.
<p>While retrieving articles with simple boolean queries will suffer from a list of problems in general, several machine learning approaches can be applied to allow better performance in terms of accuracy and efficiency. In particular, instead of naively focusing on the existence of the topic in the article, techniques with machine learning can broaden the search range by including more relevant secondary keywords in addition to the various variation of the given topic. Meanwhile, methods (such as SVM, nearest neighbor) can also be employed to prune the fetched articles to filter out those are irrelevant to our topic despite containing the exact words on the topic. Furthermore, due to the fact that articles may use different choices of words over time, time series analysis needs to be applied to keep track of the evolution of one topic during a certain period of time. In this project, I plan to design and implement a proof of concept unsupervised machine learning approach to retrieve articles and prune the results, with the assistance of mathematical models.</p>
<p>In the large majority of cases in DBpedia, it is not clear what kind of relationship exists between the entities. Instead of extracting the triples from semi-structured data only, we want to leverage information found in the entirety of a Wikipedia article, including page text. The goal of this project is to develop a framework for predicate resolution of wiki links among entities, specifically, we focus on the direct cause-effect relations between events. Our task then is to extract the cause-effect entity pairs (e.g., Peaceful_Revolution, German_reunification) from the wikipedia text. We combine the idea of using the seed data (e.g., the known cause-effect entity pairs) with training a classifier (e.g., a discriminative model--- LSTM ), so as to discover more cause-effect entity pairs from wikipedia text, which is known as distant supervised relation extraction. The procedures include the pattern matching, knowledge exploration, entity recognition, entity mapping, etc. Eventually, we aim to acquire more reliable causal relations between entities of DBpedia.</p>
<p>There is functionality for working with dynamical systems over projective space in Sage. However, one of the areas lacking in functionality is the moduli space. We say two self-maps of projective space are equivalent if there is an element of the Projective Linear Group that conjugates one to the other. The following two algorithms should be implemented, in order to create more functionality in moduli space. The first algorithm would be that given two endomorphisms of projective space determine if they are conjugate. In other words, determine if they are in the same class in the moduli space. If they are, also return the PGL element that conjugates one to the other.The second algorithm would be that given an endomorphism of projective space to compute a reduced form. This conjugation would that makes the coefficients small. There is already an algorithm implemented to return the minimal model in terms of resultant, but the coefficients can be non-optimal. The simplest approach would be to reduce the binary form describing the fixed points or if the binary form describing the fixed points is not sufficiently nice then reduce the binary form for points of period n for some small n.</p>
Kubeflow Trainer V2 currently supports only TorchTune as its LLM fine-tuning backend. TorchTune stopped adding new features in July 2025, leaving Kubeflow with ~4 supported models and no support for DPO, GRPO, or other post-training methods. This project builds a pluggable multi-backend framework for Kubeflow Trainer. The first new backend is TRL (HuggingFace's fine-tuning library), adding support for SFT, DPO, GRPO, Unsloth acceleration, and multi-node DeepSpeed. The architecture includes: an LLMBackend ABC + registry in the Python SDK for config-driven backend dispatch, a LLMFrameworkHandler interface in the Go control plane (Torch plugin) for zero-touch extensibility, and entry_points-based registration so the community can add backends like LlamaFactory without upstream changes. Deliverables: TRL backend (Python SDK + Go handler), pluggable registry, runtime configs, integration tests, and documentation. All existing TorchTune workflows remain unchanged.
<p>As a student, I use graphs in almost everything related to my master’s research. Graph Theory is a great mathematical tool for programmers and mathematicians, as it facilitates some computational problems resolution and modelling. However, is not always easy to understand the underlying theory and algorithms.</p> <p>Rocs (<a href="https://kde.org/applications/education/rocs" target="_blank">https://kde.org/applications/education/rocs</a>) is a Graph Theory IDE created in January of 2010, with new updates still being committed almost every month, currently in the KDE repositories, designed for lecturers, students and researchers. It is composed of a visual data structure and a powerful scripting engine in Javascript. Though useful for creating simple graphs, it is still lacking the tools to simplify the creation of bigger graphs, and also of basic graph theory algorithms for beginners. Another improvement it should have is a step by step execution system, in order to fully comprehend the algorithms and facilitate debugging.</p>
PersonalAIs: Generative AI Agent for Personalized Music Recommendations This project aims to enhance music streaming experiences by building an AI agent that offers emotionally intelligent, personalized music recommendations. Traditional recommendation systems often overlook users’ real-time emotional states. To address this, we propose a full-stack solution that uses large language models (LLMs), emotion detection, and Spotify API integration to tailor music suggestions based on users’ mood and preferences. Key deliverables include: • A conversational chatbot UI for user interaction • Emotion detection using LLMs • Integration with Spotify for real-time user data • A hybrid recommendation system (recall & rank) based on listening behavior and emotional context • A frontend-backend architecture for seamless user experience This project will deliver a working prototype capable of understanding user emotions and musical tastes, and recommending tracks that resonate with the user’s current mood.
The NOS OFS Skill Assessment package is a Python-based framework used to evaluate operational forecast systems (OFS) through observation retrieval, model processing, skill computation, and visualization. As the system continues to evolve, gaps in automated testing, continuous integration, and user-facing validation create challenges in maintaining reliability and usability. This project focuses on improving developer and user experience without modifying the scientific computation pipeline. The primary objectives are to introduce a GitHub Actions-based CI workflow, expand deterministic unit test coverage across core modules, and enhance the existing tkinter-based GUI with structured parameter grouping, validation logic, and contextual tooltips derived from existing argument definitions. In addition, the project will improve contributor onboarding by restructuring documentation using MkDocs. A lightweight, read-only dashboard will also be explored to organize and present outputs from completed skill assessment runs, providing a concise operational overview.
OpenSSL 3.2 introduced native QUIC support, enabling HTTP/3 on top of the world’s most widely deployed TLS library. Apache httpd, however, is still architected around TCP-based connections and its Multi-Processing Modules (MPMs), which makes integrating QUIC and HTTP/3 non-trivial. An example codebase (openssl-h3-examples) already demonstrates HTTP/3 using OpenSSL’s QUIC support and nghttp3 in a standalone server, and reuses patterns inspired by mod_http2, but it remains experimental and is not integrated into httpd as a module. This project proposes a new HTTP/3 module for Apache httpd, building on OpenSSL QUIC and nghttp3, and aims to resolve the architectural mismatch between httpd’s TCP-oriented worker model and QUIC’s UDP-based, stream-multiplexed transport. The goal is a maintainable, reviewable prototype module that the Apache HTTP Server community can evaluate and evolve toward production readiness.
My proposal is to extend the Python package dingo with two Markov Chain Monte Carlo sampling methods, which will enable exponential sampling from the set of flux vectors for constraint-based metabolic modelling. The sampling methods - NUTS Reflective Hamiltonian Monte Carlo, Reflective Exact HMC, and Riemannian Hamiltonian Monte Carlo - have already been implemented in C++ and integrated into the volesti codebase. The extended C++ bindings and Python wrappers will enable access to these methods through dingo. This will involve running experiments on benchmark metabolic networks to compare the results obtained from the new random walks and the existing Multiphase Monte Carlo Sampling method based on Billiard walk. The experiments will generate new biological insights by providing an exponential sampling of flux vectors, which will allow for an unbiased characterization of the metabolic capabilities of biochemical networks. The final deliverables of the project will be the extended dingo package, the results of the experiments, and a brief report discussing the findings.
This study focuses on the spatio-temporal dynamics of forest wildfires in Fairbanks under a changing climate. The research aims to analyze the patterns, causes, and impacts of wildfire occurrences using geospatial techniques. Satellite data such as Landsat and MODIS will be used to derive key indices including NDVI and NDMI to assess vegetation health and moisture conditions. Land Use/Land Cover (LULC), slope, aspect, wind speed, temperature, and rainfall data will be integrated within a GIS environment. Temporal analysis will be conducted to identify trends in temperature and precipitation, while change detection techniques will be applied to evaluate landscape transformations. A fire susceptibility map will be generated using a multi-criteria decision analysis approach to identify high-risk zones. The study will also examine the relationship between climate variability and wildfire frequency. The results will support early warning, risk assessment, and effective wildfire management strategies in the Fairbanks region.