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 introduces WHOIS details lookup and fuzzy location estimation to improve OpenWISP’s device management capabilities. When a device’s reports a public IP, WHOIS information will be retrieved asynchronously using Celery and the geoip2 library, ensuring non-blocking performance. Estimated locations will be marked with an is_estimated flag. Administrators will be able to filter such entries in the admin interface using custom filters, while APIs will provide equivalent filtering through django-filters. Serializers will expose both summary and detailed WHOIS records via the API. Configuration of these features will be available at both global and organizational levels using FallbackBooleanChoiceField, maintaining consistency with existing configuration practices. With this implementation, users will be able to view WHOIS data and approximate device locations when IP changes occur, refine locations manually, and receive notifications upon location updates. The Device admin UI will display WHOIS details and support filtering by estimated locations, while the Location admin will include clear indicators and filters for estimated entries. APIs will deliver the same functionality, ensuring consistent behavior across interfaces.
Reproducibility as Accuracy (RaA) is a benchmark which aims to evaluate how effective multimodal AI systems are in preserving information fidelity during iterative transformations from image to text. By implementing a recursive process—converting an image to text and back to an image over multiple iterations—the benchmark will apply qualitative and quantitative metrics to assess semantic drift and information degradation. Deliverables include an open-source benchmarking framework, a suite of evaluation metrics, documentation, and analytical reports, providing insights to enhance the reliability of multimodal AI systems.
Meshery integrates with a bunch of CNCF tools, and keeping those integrations stable needs proper automation. The goal of this project is to improve and expand end-to-end (E2E) test coverage using Playwright, so that the Meshery UI and workflows stay reliable even as new changes come in. I'll start by understanding the current test setup and structure, identify gaps or flaky tests, and then refactor or write new Playwright tests that are modular and clean. These tests will be fully integrated with the CI (GitHub Actions), ensuring they run with every PR and provide quick feedback. Deliverables include new E2E tests for major workflows, fixes for unstable tests, performance improvements through parallelization, updated documentation for contributors, and a more robust CI test pipeline. I'll be actively collaborating with the Meshery community throughout
This project developed scalable Nextflow pipelines for variant analyses across diverse eukaryotic genomes. The workflows support calling germline and somatic variants, computing key variant metrics, generating quality-control reports, and producing standardised outputs for efficient downstream analysis. Clear documentation ensures researchers of all backgrounds can run the pipelines confidently and explore genomic variation with ease.
This proposal aims to upgrade and expand existing open-source tutorials and examples to support the new unified Gemini SDKs for JavaScript/TypeScript and Python. My approach involves three key components: (i) Cookbook Migration: Migrate existing Python examples in the Gemini Cookbook to idiomatic JavaScript/TypeScript. I will ensure the code follows best practices in both languages and is thoroughly documented like the already existing cookbook examples. (ii) New Tutorials & End-to-End Examples: Develop fresh, comprehensive tutorials demonstrating a variety of Gemini use cases. Examples include building a chatbot, creating a text summarization tool, and generating images from text prompts. Each tutorial will include step-by-step instructions, code samples, and in-depth explanations of underlying Gemini concepts. Deliverables include: - A set of migrated and newly developed code examples in both Python and JavaScript/TypeScript. - Comprehensive, user-friendly tutorials and documentation that clearly explain the code and underlying concepts. - Updated examples for existing open-source libraries utilizing the Gemini SDK, with clear migration guides.
The LibreOffice BASIC IDE currently offers limited code assistance, hindering developer productivity with unreliable completion, no object browser, and no inline help (parameter/quick info). This project will implement a modern, comprehensive code assistance system. This involves creating [1] a robust backend analysis engine to parse BASIC code (editor/libraries) and introspect UNO objects for detailed symbol information (types, scopes, signatures), and [2] an efficient knowledge cache. These components will power [3] user-facing tools including: an integrated Object Browser for API exploration, significantly improved context-aware code completion (Ctrl+Space and enhanced '.'), and helpful Quick Info/Parameter Info tooltips. The final deliverable will be these features fully integrated and enabled by default, delivering a significantly more productive and intuitive BASIC development experience in LibreOffice.
Physicists at the LHC need to perform ML inference on massive amounts of data. Currently, the bookkeeping of ML model files is an unsolved problem. The goal of this project is to evaluate the CernVM File System (CVMFS) as a platform to store, organize, and distribute model files. The two primary issues to tackle are latency and infrastructure. We know there is a latency with using CVMFS versus local storage, so I will rigorously benchmark the performance of CVMFS and determine ways to minimize this overhead. For infrastructure, I will test integrating different services with CVMFS such as Cern Document Server (CDS) or Kubeflow's KServe. The final deliverables will include extensive documentation on benchmarks and best-practices for using CVMFS, and an overall evaluation of CVMFS and the possible integrations like KServe. An example deployment of ML models from the ML4EP project using CVMFS will also be demonstrated.
This project aims to use self-hosted runners to run GPU-intensive tasks like LLM blueprint or (planned) AI Playground. The necessary infra is provided by Oracle, plan is to use Oracle Kubernetes Engine (OKE) with NVIDIA GPUs for this task. Any code or sample that requires GPU-intensive resources will be transferred to OKE infra instead of generic GitHub infra for faster and more efficient execution. For now, the idea is to have a specific policy that whenever any Jupyter Notebook code will be added to a trainer/examples/pytorch/** folder (e.g., in trainer/example/pytorch/image-generation/sample.ipynb), that action is transferred to OKE infra by the GitHub self runner. For security reasons, this process will require manual approval from one of the maintainers to trigger the self-runner build. I will set up the GitHub workflow to monitor changes in the respective folder. Once approved, the CI action will execute the code using the GitHub self-runner on the OKE infrastructure. Additionally, we will set up a dashboard for monitoring and metrics to understand usage patterns and identify bottlenecks. The scope of this project is set up on OKE, but theoretically, this is platform-agnostic; it can be deployed on any Kubernetes cluster with sufficient GPU resources.
This project addresses the lack of automated testing in InVesalius, which currently relies on manual verification, slowing development and risking regressions. By implementing a comprehensive pytest-based test suite (unit, functional, and GUI tests) and integrating GitHub Actions CI/CD, the project will ensure core functionalities like DICOM processing, 3D visualization, and GUI interactions remain stable as new features are added. Deliverables include: 1)A structured test suite covering the entire inVesalius codebase. 2)GitHub Actions workflows to automate testing on commits/pull requests. 3)Mock file systems and dependency isolation for safe, repeatable tests. 4)Well-documented test cases to guide future contributors. This will reduce reviewer workload, accelerate debugging, and foster community growth by lowering the barrier to entry for new developers.
This project aims to build an automated workflow to register BIDS-compliant neuroimaging datasets from external repositories (e.g. Public nEUro) into the EBRAINS Knowledge Graph. Using the bids2openMINDS converter and EBRAINS APIs, the pipeline will convert BIDS metadata into openMINDS JSON-LD format and programmatically submit it to EBRAINS. The solution will eliminate manual submission, improving scalability, metadata consistency, and dataset discoverability. Deliverables include a secure authentication module, an end-to-end ingestion pipeline, documentation, and a user tutorial. Optional features include batch submission support and broader modality coverage.
This proposal aims to implement the Menu, Sidebar, and Floating Action Button (FAB) components in iOS for MIT App Inventor, ensuring feature parity with Android as introduced in PR #2299. The Menu Component will introduce a dropdown-style menu supporting nested items, icons, and keyboard navigation. The Sidebar Component will provide a sliding navigation panel with swipe gestures, tap-to-close functionality, and essential navigation links like Home, Profile, Settings, and Logout. The Floating Action Button (FAB) will be a circular, floating button positioned at the bottom-right, supporting custom actions, animations, and a shadow effect for a modern UI experience. A key aspect of this project is ensuring feature parity with Android, maintaining UI/UX consistency across both platforms. I chose this project because it bridges the gap between iOS and Android, ensuring that MIT App Inventor provides a seamless cross-platform experience. It also enhances user experience by introducing smooth, intuitive navigation and interactions, making iOS-built apps more functional and engaging. Additionally, this project will expand MIT App Inventor’s capabilities by adding essential missing features, making the platform more versatile and user-friendly. The estimated timeline for this project is 175 hours, covering implementation, testing, and final integration. Upon completion, this update will significantly enhance MIT App Inventor’s iOS version, making it a fully-featured, cross-platform development tool that aligns with modern app-building standards.
This project aims to develop a prototype that simplifies data exchange within Slack, enabling seamless information retrieval and input with minimal effort. The primary focus is on creating an intuitive, efficient interface to streamline workflows and enhance productivity. Additionally, the project involves researching the feasibility of extending support to other communication platforms (e.g., Microsoft Teams, Discord) by analyzing API complexities, authentication mechanisms, and integration challenges. The outcome will include a functional Slack prototype alongside a comparative assessment of cross-platform compatibility, identifying key barriers and potential solutions for future expansion. This work will contribute to a more interconnected and automated workspace ecosystem, reducing manual effort in information management.
This project aims to implement the Fourier Neural Operator (FNO) in DeepChem, expanding its support for solving complex partial differential equations (PDEs) in scientific machine learning. Traditional neural networks often struggle with PDEs due to their reliance on fixed-grid discretization. FNO addresses this by operating in the frequency domain, learning resolution-invariant mappings between function spaces using the Fourier transform, making it both scalable and efficient. The project involves developing a new SpectralConv layer and a high-level FourierNeuralOperator class compatible with DeepChem’s PyTorch backend. The implementation will be modular, enabling future extensions such as Physics-Informed Neural Operators (PINO). Deliverables include the FNO model, unit tests, and a comprehensive tutorial.
The goal of this project is to improve the performance of Blender’s sculpting system, particularly when working with high-resolution meshes. Sculpting can become unresponsive on dense geometry, especially on lower-end hardware, so this effort focuses on optimizing performance at a low level. Key areas of development include spatially sorting mesh vertex data to enhance cache coherence, implementing smarter undo data storage through compression techniques, and introducing multithreading within nodes to make better use of CPU resources. The intended outcome is a noticeably more responsive sculpting experience for artists, with faster interaction and improved overall performance.
This project will make the addition and sharing of third-party battery models easier in PyBaMM. It is an extension of the PyBaMM cookiecutter project of year 2024 and will include a Dispatch API, a single location to register models, and enhance model loading using entry points. The copier template will also be enhanced so that individuals with varying levels of experience will be able to use it more easily. By incorporating features such as dynamic model loading, caching, and enhanced documentation, the project will make it easier for more users and developers to work with external models and make PyBaMM more flexible and convenient to use in the long term.
Expanded and enhanced nx-parallel, a parallel backend for NetworkX that leverages joblib to parallelize graph algorithms. I implemented several embarrassingly parallel algorithms, analyzed their performance speedups, and optimized existing implementations with the help of a mem-mapping approach. Additionally, I switched the default configuration to use NetworkX’s system in place of joblib’s, developed a more reliable timing script to reduce measurement variability, introduced a `should_run` mechanism to determine when the parallel backend should be applied, and improved ASV benchmarks by adding setup functions for consistent performance testing. The entire development workflow was documented in a series of eight GSoC blogs, published biweekly throughout the project duration.
This project aims to fine-tune Vision-Language Models (VLMs) for domain-specific tasks like object detection and hierarchical classification, focusing on real-world applications in fields such as medical diagnostics and biodiversity monitoring. Although VLMs like BLIP-2 and TinyCLIP show strong general-purpose performance, their effectiveness diminishes in specialized domains due to domain shifts, hierarchical label structures, and hardware constraints. To address this, we will implement efficient fine-tuning strategies—including LoRA, QLoRA, and adapter tuning—within the OpenVINO Training Extensions (OTX), enabling scalable training on consumer-grade GPUs. We’ll enhance these models with support for structured prediction and deploy them using OpenVINO’s quantization and optimization toolchain, ensuring fast, lightweight inference across devices. Beyond technical contributions, the project delivers practical tools and open-source pipelines that bring advanced multimodal AI to resource-limited environments. By supporting transparent, efficient deployment of VLMs in impactful domains, this work bridges cutting-edge research with high-value, real-world use cases.
RoboStream is a comprehensive solution designed to address the challenge of monitoring and visualizing complex robotic telemetry in real time. As robots become increasingly integral to various industries, the need for an intuitive, immersive interface to track sensor data and live video streams grows ever more critical. Our project integrates advanced simulation tools and real robots with a responsive, cross-platform mobile application built in Flutter. This combination lets us capture, process, and transmit detailed telemetry directly into the immersive Liquid Galaxy multi-screen environment.
I will be developing DashBot - the AI assistant for API Dash which supercharges developer productivity by helping developers automate tedious tasks, follow best practices, interact & obtain contextual suggestions, all via natural-language input. DashBot will be designed in a modular and extensible manner and provide the following list of features: 1. Explain responses & identify any discrepancy 2. Debug requests based on Status codes & Error messages 3. Generate API documentation 4. Understand API and generate tests 5. Generate plots & visualizations for API responses along with ability to customize 6. Generate API integration frontend code for frontend frameworks like React, Flutter, etc. For each of the tasks benchmark evaluations will also be done so that it is easier for end users to choose the right backend LLM. Add support for various API Authentication Methods such as, Basic authentication, API Auth, API key, Bearer token, JWT Bearer, Digest Auth, OAuth 1.0, OAuth 2.0.
I propose to make an AI-powered tool that can analyze complex communication data from driving simulator experiments at the Translational Research for Injury Prevention (TRIP) Laboratory. The tool will process high-fidelity audiovisual data from six participants working together in networked simulators. Building on my previous work in audio extraction, transcription, sentiment analysis, and preliminary visualization, I want to create a complete solution that includes strong data pipelines, advanced NLP techniques (such as speaker diarization, NER, and topic modeling), dynamic reporting with interactive visualizations (heat maps, network graphs), and an easy-to-use user interface. The final product will be a fully functional prototype with full documentation that will help researchers learn more about how groups work together, how people talk to each other, and how drivers act in order to improve safety and inform the design of smart systems.
I will build an accessible AI support agent for Zendalona product.. This project addresses a critical gap in accessibility support by creating a conversational interface that seamlessly accepts both voice and text input, allowing users to interact naturally with Zendalona's product ecosystem. The technical architecture employs an Agentic Retrieval-Augmented Generation (RAG) approach using LlamaIndex. This system will methodically index all Zendalona product information, documentation, and support materials in a vector database (ChromaDB or FAISS). When users submit queries, the system retrieves contextually relevant information and passes it to a large language model alongside the query, enabling precise, helpful responses based on accurate product knowledge. The implementation follows a three-tier approach: First, a comprehensive data ingestion pipeline that processes, cleans, and vectorizes Zendalona's product information. Second, a robust Python backend built with FastAPI that handles query processing, context retrieval, and response generation through integration with STT/TTS services and LLM APIs. Third, a fully accessible React-based multimodal chat interface featuring high-contrast text, complete keyboard navigation, ARIA attributes, and seamless screen reader compatibility, particularly with Orca. Key deliverables include a modular, containerized system with separate frontend and backend Docker containers orchestrated via Docker Compose, comprehensive API documentation, extensive accessibility testing results, and detailed user guides. The solution prioritizes WCAG 2.2 compliance, minimal latency, data privacy, and security throughout the architecture. The final product will provide visually impaired users with an intuitive, conversational way to access product support that works effectively with their existing assistive technologies.
This project aims to decode the neural dynamics of social interactions by adapting the CEBRA framework to analyze dyadic EEG data. Focusing on differences between neurotypical individuals and those with autism, the work will develop robust preprocessing pipelines, contrastive learning methods for neural synchrony, and interpretable clinical biomarkers. The deliverables include an open-source toolkit for EEG analysis, enabling researchers to study brain interactions with greater precision. This approach could advance early diagnosis and deepen our understanding of social brain function.
A majority of the Gemma chat applications on Hugging Face Spaces do not allow the user to adjust generation settings or system prompts, giving the user only a generic experience. This project aims to provide users with total control of the model. Users will be able to fine-tune settings such as temperature, top-p, repetition penalty, and max tokens, allowing for a more personalized experience. Additionally, the project will introduce a range of predefined model personalities as well as custom context, enabling users to choose from different response styles, tones, and behaviours that best suit their preferences and use cases. I will deliver an open-source web application with comprehensive documentation accompanying it, which includes instructions, usage guides, and explanations of how different parameters influence the model's output. I believe that this project will be a valuable addition to the Gemma ecosystem on Hugging Face.
The current app experience can be frustrating for learners who are redirected to earlier concepts for revision after making a mistake, as they must re-answer all intermediate questions to return to their original point. This proposal introduces a new Flashback feature that allows learners to view the correct answer or revisit previously solved question answer tied to their misconception and then return directly to the point where they became stuck—eliminating the need to reattempt intermediate steps. This reinforces foundational understanding while enabling a smoother, frustration-free learning journey. Additionally, this project improves the visual presentation of submitted answers for multiple choice and item selection interactions by replacing the current HTML-based approach with a cleaner, natively rendered experience.