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.
This project addresses critical gaps in Gemini API integration across leading agent frameworks (LangChain, LlamaIndex, CrewAI, Composio), where multimodal capabilities and function calling remain underdeveloped or inconsistent. By implementing standardized components for multimodal processing, function calling, and performance optimization, the project will democratize access to Gemini's advanced capabilities for developers building sophisticated AI agents. The implementation will follow a systematic approach, beginning with a comprehensive framework audit and gap analysis, followed by development of core multimodal support, function calling capabilities, and performance optimization layers. The solution will include a unified adapter layer that standardizes Gemini API access while respecting each framework's architectural patterns. Deliverables include: (1) enhanced framework integrations with full multimodal support, (2) standardized function calling implementations, (3) performance optimization components including token management and caching systems, (4) comprehensive documentation and examples, and (5) a benchmarking suite for performance analysis. This project will establish a new standard for Gemini integration in the agent ecosystem, enabling entirely new classes of multimodal AI applications.
Currently, the Cookbook has multiple resources that can help any event became inclusive and accessible to everyone. However, its accessibility remains limited due to its exclusive support for English. This language restriction creates a barrier for a global audience, preventing many from fully benefiting from its guidance. This project aims to develop a comprehensive translation workflow for the DISCOVER Cookbook, enabling seamless multilingual support. A language selection mechanism will be integrated within the book, allowing users to easily switch between available translations. Additionally, an automated update system will be implemented to ensure translations remain synchronized as content evolves. To maintain high-quality translations, the project will engage and onboard community translators, fostering collaboration and sustainability in the translation process.
This project aims to improve Jenkins CI/CD troubleshooting by creating a domain-specific Large Language Model (LLM). The core task involves fine-tuning an open-source LLM (e.g., Phi-4/Qwen 2.5) on structured data derived from ci.jenkins.io build, logs and metadata, using resource-efficient techniques like QLoRA with Unsloth. To ensure feasibility despite potential API limitations, a primary strategy involves heuristic/manual data curation, supplemented by LLM assistance where possible. A baseline Retrieval-Augmented Generation (RAG) system incorporating Jenkins documentation will be implemented to enhance factual accuracy. Key deliverables include the fine-tuned model (GGUF for Ollama), a basic FastAPI backend, the RAG pipeline, and an evaluation report incorporating both automated metrics and qualitative mentor feedback
I propose revamping the PSLab.io website to give it a clean, modern, and professional look while ensuring long-term maintainability. The project will involve migrating the site from Elementor to Gutenberg, streamlining the editing experience and making it easier for contributors to update and manage content. By focusing on modern UI/UX design, improved performance, accessibility, and SEO optimization, the redesigned site will better highlight PSLab’s open-source hardware and software initiatives. This transition will not only improve usability for visitors but also simplify content management for the community, ensuring a sustainable and future-proof digital presence.
The latest CUPS 3.x versions support only driverless printing through the new IPP Everywhere architecture. Modern printers predominantly use driverless technology, supporting IPP printing and automatically configuring themselves via CUPS. Unlike traditional setups, these IPP printers do not require a permanent print queue. However, many printers still rely on drivers (PPD + filter). To address this, the Genome Control Center (GCC) must be adapted to support both driver-based and driverless printers. While significant progress has been made, substantial work remains before this new approach can be fully integrated into GCC. The project aims to complete all remaining tasks, ensuring a seamless user experience for GCC printers while refining all aspects for the final release.
This project will involve refactoring the Human Neocortical Neurosolver (HNN-core) codebase to enable flexible development of neural circuits with customizable cell types. Currently, HNN-core contains hard-coded assumptions about cell types, locations, and connectivity that limit researchers' ability to implement networks with different architectures. I will restructure key modules to replace these hard-coded constraints with dynamic definitions, allowing users to flexibly select and connect cell types with different morphologies and electrophysiological properties. The solution involves refactoring cells_default.py for dynamic cell creation, modifying network.py to handle arbitrary cell types, and updating related components to support custom cell types. The proposed plan includes dynamic cell type implementation, updated network connectivity handling, a demonstration model with custom cell types, comprehensive documentation, and unit tests ensuring backward compatibility. This work will significantly enhance HNN's capabilities for hypothesis testing in EEG/MEG research while maintaining its user-friendly approach for neuroscientists.
JSON Schema validation errors are often cryptic and difficult to understand, making debugging frustrating for developers and end users. This project aims to develop a JavaScript library that transforms raw JSON Schema validation outputs into clear, concise, and human-friendly error messages. It will maintain compatibility with the standard format while improving error readability, supporting multiple languages, and allowing customization. By the end of the project, developers using standard-compliant JSON Schema validators will have access to a high-quality, user-centric Better JSON Schema errors—enhancing the developer experience and making debugging more intuitive for end users.
The GNU Octave Statistics package lacks class definitions for knnsearch and rangesearch, and its KDTree method is currently disabled due to poor performance (GitHub issue #151). This project aims to implement KDTreeSearcher, ExhaustiveSearcher, and hnswSearcher classes, along with the createns helper function, to enhance clustering capabilities. The KDTree implementation will be optimized using a compiled Octave (.oct) library for faster index construction and queries. Each searcher class will include efficient knnsearch and rangesearch methods. Deliverables include optimized implementations, extensive testing, benchmarks, and comprehensive documentation, improving performance, usability, and MATLAB compatibility.
There are many static .mdx and .json files in the codebase that are used to render pages or components on the AsyncAPI website. These static files are regularly updated by some scripts on builds and via cron jobs and my task is to refactor the scripts. Currently, some scripts contain a main function that directly executes the scripts. In this refactoring, I am removing those direct executions and instead call all script functions from a central npm folder. This will allow us to structure integration tests directly for these scripts. As a next step, I will be writing the integration tests as well for these script and fix the existing unit tests affected due to refactoring. Tech Stack: TypeScript, Data Structures and Algorithms, NPM, Testing, Unit Testing, Integration Testing, Code Refactoring, Next.Js
The OpenELIS organization currently relies on JSP (Jakarta Server Pages) for most of its pages but is gradually transitioning to a modern JavaScript-based framework, React.js. Many pages still use Java-based controllers and endpoints to dynamically update and display data on the front end. As part of this project, I will rewrite all Test Management Pages and their components in React UI, integrating them with REST API endpoints to fetch and render data. Additionally, I will migrate all Test Management Configurations, Management, and other operational pages to React.
The project aims to develop awar s companion app for the Ultimate Alarm Clock mobile application. This companion app will allow users to sync, control and manage alarms directly from the smart watch using the Data Layer API, also lets them set alarms natively on watch and that message can be send to UAC as well.
The Active Segmentation Platform for ImageJ (ASP/IJ) aims to provide a general-purpose workbench that allows biologists and other domain experts to access state-of-the-art techniques in machine learning to achieve excellent image segmentation and classification. This project aims to speed up the convolution engine in ASP/IJ by implementing GPU parallelization using JCuda, a Java-based framework for CUDA. Currently, the convolution engine runs only on the CPU.
This project evaluates large language models on frontend coding and UI generation tasks by extending and integrating ideas from open-source coding benchmarks such as Multi-SWE-bench and WebDev Arena. The evaluation framework assesses coding and UI generation capabilities across various dimensions, with particular emphasis on multimodal features for frontend development. It leverages the models’ ability to generate, render, and iteratively refine solutions using visual feedback from rendered UIs, and supports multiple frontend frameworks (React, Vue, Angular, Svelte, Next.js) alongside single-file HTML tasks. The solution involves building a modular pipeline with typed configuration, provider adapters (Ollama, vLLM, OpenRouter), and a model manager; implementing multimodal evaluation loops that incorporate screenshots and structured judging; supporting iterative evaluation (allowing models to refine solutions over multiple attempts); expanding the task taxonomy with increasingly complex frontend challenges (calculators, dashboards, interactive apps, framework-based projects); and automating the entire evaluation pipeline with reproducible runs and detailed artifact logging. Deliverables include the enhanced evaluation framework with new benchmark tasks, structured outputs and reports (JSON, screenshots, summaries), automation scripts for reproducible runs, and thorough documentation. The framework provides a transparent and extensible way to compare advanced AI coding capabilities in frontend development without relying on a leaderboard, focusing instead on rigorous, reproducible benchmarking.
The Robotics Simulation for Liquid Galaxy 2025 Project is an innovative addition to the ever growing Liquid Galaxy Ecosystem. This project aims to combine multi-display synchronous visualization with advanced robotics simulation which is aimed to be controllable through a smartphone. This project bridges the powerful networking and visualization capabilities of Liquid Galaxy with cutting-edge robotic technologies such as ROS (Robot Operating System) and Gazebo, a popular simulation tool for testing and developing robotic systems in virtual environments.
This project proposes to improve Mifos X's loan origination capabilities by introducing a flexible, plugin-based architecture for integrating credit bureaus. At present, the system supports only a single provider and is tightly coupled, which makes it difficult to extend. The proposed work aims to address this by building a generalized plugin interface, enabling integration with multiple providers, and incorporating a FICO-based credit scoring mechanism. Additionally, I hope to enhance the user experience through new interfaces that support better credit insights during loan processing. The intention is to contribute a reusable and maintainable solution that aligns with Mifos X’s mission of accessible and responsible financial services.
This project aims to enhance Kubeflow's security posture by implementing Istio CNI by default and providing an optional Ambient Mesh configuration. The work involves securing the service mesh, creating controllers for HTTPRoute and AuthorizationPolicies, updating manifests to support these new features, securing KServe with proper authentication protocols, and making Kubeflow rootless by default. The project will also introduce performance benchmarking so that users can compare between the three available modes: Istio CNI (default), Istio init containers, and Ambient Mesh.
This project will add bitmap convolution support to Doodle, enabling powerful image processing effects like blurring, sharpening, edge detection, and embossing. The implementation will design a clean functional API for convolutions that works across Doodle's backends (Java2D and SVG), despite their different underlying implementations. For Java2D, the solution will render elements to bitmap buffers before applying convolutions, while for SVG, it will leverage native filter capabilities. The project will include standard convolution kernels, composition operations, comprehensive documentation, and examples. This enhancement will significantly expand Doodle's capabilities beyond vector graphics while maintaining its functional programming model.
This project focuses on developing a new machine learning approach called Quantum Model-Agnostic Meta-Learning (Q-MAML) specifically for high-energy physics (HEP) experiments like those at the Large Hadron Collider (LHC). These experiments produce massive amounts of data that change over time, making it challenging to train machine learning models effectively. Q-MAML uses quantum computing techniques with variational quantum circuits to enable "few-shot learning." This means the models can quickly adapt to new data with very little retraining, which is crucial for the dynamic environment of particle colliders. The goal is to improve tasks like identifying particle collision patterns (event topology classification) and finding rare signals. The project will develop a modular, open-source Q-MAML workflow that is compatible with both quantum simulators and actual quantum hardware, utilizing platforms like Qiskit and PennyLane. The performance of Q-MAML will be rigorously benchmarked against classical MAML and standard variational quantum algorithm (VQA) classifiers on physics-inspired tasks. These benchmarks will include detailed evaluations of training curves, adaptation behaviors, and resource analysis. The deliverables will consist of a fully documented codebase, interactive Jupyter notebooks for ease of use, and reproducible benchmarking tools. This work aims to establish a practical foundation for hybrid quantum-classical learning in scientific applications, particularly within the field of high-energy physics.
This project aims to enhance Blender’s Glare node by addressing UX and technical limitations, improving usability, and performance. By resolving long-standing issues such as unintuitive controls, resolution dependency, and bad performance particularly in modes like Simple Star and Streaks. The updated Glare node will offer a more intuitive and efficient implementation. The general improvements I aim to implement are stated in the following task: Issue: Compositor: Future of Glare node #124176
This project aims to build an AI-powered GitHub assistant that helps open-source maintainers review code, detect security issues, and prioritize tasks efficiently. The assistant will analyze pull requests, identify bugs and vulnerabilities, and provide real-time feedback using LLMs and advanced rule-based scanning. A smart prioritization system will rank issues based on urgency and impact, ensuring maintainers focus on the most critical work. The project will deliver a GitHub-integrated bot, an AI-driven code review system, and a dashboard for managing tasks, making open-source development faster, more secure, and easier to maintain.
The Jenkins website currently relies on Awestruct, an outdated static site generator, leading to maintenance challenges, inconsistent user experience, and missing versioned documentation. To address this, the project aims to migrate non-versioned content to Gatsby and versioned documentation to Antora. The solution includes modernising the UI/UX with shared React components, integrating Algolia search, maintaining SEO-friendly redirects, and automating builds for smoother contributions. Key deliverables include a full content migration, unified responsive design with a Jenkins theme, cross-version navigation, and a CI/CD pipeline for seamless deployments.
About 88 million U.S. households now identify as campers, and ~54 million households took a trip last year National parks run thousands of search-and-rescue operations every year (e.g., ~3,400 in 2022), and research shows tens of thousands of people were lost in parks over the 2004–2014 decade. First aid, navigation, shelter, and water skills reduce risk and help you protect others. That’s why I built Gemma Scout: an on-device, privacy-first AI assistant built for wilderness survival and outdoor adventures. Powered by a finetuned version of Gemma 3N, it combines domain-specific survival knowledge with real-time multimodal reasoning to guide users through critical tasks like plant identification, mushroom safety, shelter building, and first aid. The model was trained on over 30 million tokens and 7,618 curated Q&A pairs, sourced from expert handbooks and image datasets of edible plants and mushrooms. Leveraging Unsloth’s optimized finetuning framework with LoRA and quantization, Gemma Scout runs efficiently on mobile devices without internet access. Using LLM evaluations, Gemma Scout outperformed its base model in survival QA by a win rate of 89%, and achieved over 36% improvement in mushroom classification and 47% in plant identification. It supports multimodal chat via llama.cpp mtmd and Swift integration, enabling users to send images and receive grounded, expert-level responses—completely serverless. Whether you're camping, hiking, or lost in the wild, Gemma Scout helps you make smart, safe decisions—even when you're off the grid.
This project aims to enhance Artemis, a modular vulnerability scanner, by automating technology-specific scans, expanding vulnerability detection, and improving usability. Key tasks include implementing Nuclei workflows, adding support for detecting new vulnerabilities, creating a generic login bruteforcing tool, enabling direct URL scanning, and optimizing performance. Additionally, I will actively address critical issues to ensure Artemis remains robust and efficient.
Oppia's Exploration Player has several issues affecting the learning experience. Learners struggle to distinguish between hints and solutions, and the audio player is hard to find on desktop and takes up too much space on mobile, reducing content visibility. Progress tracking is unclear, and navigation between lessons is difficult. On mobile, limited space and a cluttered UI make it challenging to focus on the content. This proposal aims to redesign the player for better clarity, usability, and efficiency, ensuring a smoother experience across all platforms.