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.
<p>In this proposal, I describe my plan to adapt Gatekeeper’s priority to the Linux kernel in context of traffic control and QoS. Gatekeeper is an open source defence against denial-of-service (DoS) attacks. To protect server resources from an attack, it services requests based on priority: the higher its priority, the closer to the exit of the egress queue its placed. Additionally, it also limits the requests’ maximum link bandwidth occupancy to 5%. Such countermeasures are implemented through the priority queue data structure which holds all requests in a single priority queue dropping low-priority requests when resources are low. This functionality allows us to allocate as many resources as possible to high-priority requests while servicing low-priority requests only when resources are idle. The successful implementation of this project will essentially enable a regular Linux router to help isolate and protect legitimate traffic during DoS attacks by employing Gatekeeper’s priority queue as a kernel module, enabled by its userspace component in Linux's tc(8) utility.</p>
<p>A changepoint is typically defined as a point in time where the distribution of a data-stream changes in a distinct manner, for example, typically one may look for changepoints in mean, and/or variance. Usually, this is performed in an unsupervised setting where we have no labelled examples of true changepoints. However, in practice, we usually have examples of periods of time where we know no changes should be present, or conversely where changes are expected to exist. When and where such information is available, we can potentially use this to aid our judgement of how to set complexity penalties in the changepoint estimation task, and thus, decide on an appropriate number of changepoints, a task which currently requires time-consuming parameter tuning by domain experts.</p> <p>In this particular project, we seek a simple visualization interface to support, i) human labeling of the data with the aid of several complementary measures on the data-stream such as as mean, trends, min, max, variance, etc. and ii) interactive exploration of the result space suggested by a changepoint detection algorithm. Eventually, any visualization will help communicating the data and respective decisions.</p>
My project for GSoC 2025 focuses on improving the deployment process of Apache Airavata by containerizing its core services using Docker. Currently, Airavata services are packaged as Java bundles, which makes deployment complex, inconsistent across environments, and difficult to manage. My goal is to create Dockerfiles for each service and set up a complete Docker Compose environment for easy local development and testing. I will also integrate CI/CD pipelines using GitHub Actions or Jenkins to automate the build and deployment processes. This will not only simplify dependency management and make deployments more reliable but also improve the overall development experience for contributors. Throughout the project, I’ll optimize the Docker images for production use, support different configurations for development and production, and provide full documentation along with a final demo. With my background in DevOps, containerization, and open-source contributions including work on MetaCall, OpenMRS, Apache Hadoop, and Airavata itself I’m confident in my ability to deliver a high-quality solution that will benefit the Apache community.
<p>PiType is an already existing protein-protein interaction classifying tool created by my mentor. It currently exists as both a web app and a standalone software package. PiType is designed to classify proteins within a network into different categories: Mutually Exclusive (ME) vs Simultaneously Possible (SP) and Obligate (O) vs Non-Obligate (NO). To do this, PiType takes a series of proteins within a network, analyses them to determine which of the above categories they fit into, and outputs that information as a table. Because of the current input and output format, it is difficult for biologists to analyze and interpret the data. Cytoscape is open source software that provides numerous modules for studying and visualizing protein networks, and as such is the perfect tool to use for PiType. My project over the summer will thus be to create a Cytoscape plugin app using PiType’s already existent algorithms, to create a visualization tool that provides an intuitive interface for biologists. By taking the base features of PiType, and implementing them as a Cytoscape plugin, my project will widen the number of potential users, and the utility of the app.</p>
FreeCAD is a powerful CAD software with vast capabilities supporting complete technical design pipelines beginning from modelling to FEM analysis to 3D printing. By enabling FreeCAD users to perform multibody dynamic (MbD) calculations, convert the results to FEM loads and visualise the stress and strain of an assembly in motion in a single coherent workflow, FreeCAD will be better positioned to rival commercial CAD software used by the industry today. The goal of this project is to help build that workflow by improving the combined Motion workbench so that motion-derived loads can be transferred into FEM analyses in a structured and user-friendly way. This project will focus on three main objectives. The first objective is to organise the current code into a clearer object-oriented framework. The second objective is to convert idealised MbD joint forces and torques into realistic distributed FEM loads on selected contact surfaces of real parts. The third objective is to design a user interface (UI) that allows the selection of the contact regions in FreeCAD.
This project addresses the need for enhanced usability and learning resources within the toqito quantum information library by creating a user-friendly example gallery, inspired by successful models like scikit-learn's. The core problem is the current lack of diverse, executable examples demonstrating toqito's capabilities in practical scenarios. The proposed solution involves developing the gallery's web infrastructure (potentially using Sphinx-Gallery) integrated into toqito's existing documentation, and populating it with high-quality, self-contained Python examples. Key deliverables include the functional gallery structure and 6 core examples demonstrating crucial toqito functionalities: modeling cryptographic protocols (Bit Commitment, QKD QBER) using ExtendedNonlocalGame, verifying state antidistinguishability thresholds from recent literature using the state_opt module, and illustrating channel properties (is_extremal) using channel_props. This gallery will significantly improve toqito's accessibility for researchers and new users, providing practical code and explanations for core quantum information concepts.
<p>This project aims at improving the workflow for managing streams by improving their editing UI, adding more permission sets, refactor existing permissions to make it is easy for both users and administrators to handle subscriptions for streams more easily by finishing long-pending features like Default Stream Groups.</p> <p>As stream settings and left sidebar is closely affected by any change in either of them so a parallel aim would be to improve Zulip's left-sidebar according to changed stream settings features and also to add new features to it so that it becomes more interactive rather than just listing streams/topics/PMs.</p> <p>I also aim to improve the performance of web app for narrowing to different message views, by fixing various high priority issues that are related to the message view area. Along with performance-related issues in message viewing I aim to solve issues related to interleaved views in Private message views by utilizing recent topics UI for this and improving image previewing support for messages.</p>
Hyperledger Iroha 1 has a great set of commands and queries, but some useful commands and queries are missing. Iroha 1 blockchain can be the core of banking systems because it supports transactions, multiple currencies, etc. Nevertheless, advanced filtering of transactions is only partially available in Iroha 1, and it's necessary for banking systems. Currently, to filter transactions, we need to return many of them and filter with our code - it is much heavier for both Iroha's node and client application. During the internship project, queries should be extended with optional arguments. The mentee will be able to learn the following: - the architecture of Iroha (1.x), - work in the true spirit of open-source, communicating with the Iroha community, joining calls, etc. - writing documentation, - following rules of the open-source code of Hyperledger Iroha - creating automatic tests of code Expected Outcome - Implementation according to description and plan of documentation and tests. - Added the possibility of running modified queries from Iroha's client libraries. - Present changes during the Iroha bi-weekly meeting at the end of the internship.
<p>Ribosomally synthesised and post-translationally modified peptides (RiPPs) are a class of secondary metabolites that share a common biosynthetic origin - a precursor gene, encoded on the genome of a microorganism. It is translated to a peptide by the ribosome, and then further modified to give the active final product. RiPPs can be divided in more than 20 sub-classes and have a wide range of biological functions: among others, they can work as antibiotics, nutritional additives or food preservatives. Thus, the use of RiPPs as drugs is increasing and several RiPP derivatives are already in advanced phases of clinical trials. In this project, prediction modules for a number of RiPP clusters will be developed. Working together with the mentors, who can provide technical and scientific advice, antiSMASH plug-ins will be created to allow researchers to identify the RiPP precursor gene, identify the cleavage sites of the core peptide and predict further modifications to its sctructure. With these plug-ins, we hope to accelerate all the research involving RiPPs and consequently, shorten the time “between the bench and the bed” for possible new drugs.</p>
As multimodal agents become increasingly integrated into real-world applications, ensuring their safe and reliable tool-use behavior is paramount. We introduce ATIA, a novel evaluation suite designed to systematically assess the vulnerability of tool-calling agents to adversarial inputs. ATIA focuses on crafting malicious multimodal inputs—combinations of images, text, and other modalities—that covertly manipulate an agent's decision-making process regarding external tool invocations. By simulating a variety of attack vectors, including visual prompt injections, image–text conflicts, and spurious cross-modal cues, ATIA quantifies critical metrics such as tool call correctness, unsafe tool invocation rate, and adversarial success rate. This benchmark not only exposes potential weaknesses in current multimodal agent architectures but also provides actionable insights for developing robust, safe, and aligned systems. ATIA aims to serve as a reproducible and extensible framework, driving forward research in adversarial robustness and security for next-generation AI agents.
Virtiofs is a shared file system that lets virtual machines access a directory tree on the host. Unlike existing approaches, it is designed to offer local file system semantics and performance. Currently, virtiofsd integrates the sandboxing code and the server code in a single binary. The goal is to extract that code and create an external tool that creates a sandbox environment and runs virtiofsd in it. This will make it easier to develop sandboxing features and virtiofsd side by side. In addition, that tool should be extended to be able to run virtiofsd in a restricted environment with Landlock. Landlock is a Linux kernel module that allows unprivileged processes to restrict their own access rights. This will enable the tool to apply access restrictions to virtiofsd without requiring privileged access, in a way that is enforced by the kernel, thereby increasing the protection of virtiofsd users. This will allow greater flexibility when integrating virtiofsd into a VMM or running it inside a container, due to how Landlock enforces access restriction rules as composable security layers that are compatible with system access controls.
I plan to create an R package that extends and adds additional functionality to the "You Draw It" tool. To achieve my goal I will use an iterative workflow, and prioritize tasks based on size and importance. First, in June, I will package the "You Draw It" tool for flexible usage with any real or simulated one-to-one data set. I will slowly implement features, create tests, and write documentation for each feature. In July, my next task will be to add a feature to the "youdrawit" tool that will automatically set up scales based on the range of the provided data. This function will analyze the data and determine minimum and maximum values for each axis to set up the axes' scales. In August, to improve the user experience of the "You Draw It" tool, I will add tool tips to the shiny applet that will explain tasks and help educate users on using the tool. If I can complete these tasks during the 3-month timeline and have enough time, I will move on to the two additional optional tasks. Then, I will finalize all vignettes, documentation, and tests.
The lack of automated testing capabilities in the cloud infrastructure of a data-intensive system significantly increases the time and effort required to ensure application quality. This is particularly cumbersome when the Quality Assurance (QA) team needs to set up its testing environments, connect to the development environment, and then execute test plans. A well-established CI/CD pipeline would support quality assurance and integration efforts, thereby streamlining the process. The objective of this project is to construct a CI/CD pipeline that automates the test suite execution process and improve test coverage comprehensively. This includes automated tasks such as adding authentication, installing dependencies, and configuring integration test scenarios, which are currently performed manually. By enhancing CI/CD and automating testing procedures, it will significantly reduce the team's time investment in the validation phase of the development cycle for the cloud system. This not only ensures the reliability and accuracy of D4CG platform’s data but also enables researchers to utilize this data to advance cancer research and increase opportunities for discovery.
<p>Many less resourced languages don’t have a tagged corpus or only have a small amount of poorly tagged material. In such cases unsupervised learning is a great first step beyond simply picking a random tag or the most frequent form according to some corpus statistics. The way CG is currently combined with the tagger works reasonably well empirically but is actually somewhat unsound in terms of the theory of operation of the tagger (words are identified by different ambiguity classes during training and tagging). For my first task, I will implement new ways of combining them both during in tagging and during tagger training .</p> <p>For my second task I will implement an averaged perception POS tagger based on the one in the Python module nltk. This will have some level of configurability in terms of word features (ie can choose to look at suffixes or prefixes of different lengths) to accommodate different types of languages. Another requirement will be to work (not necessarily very well) with little or no configuration (which means eventually not using the coarse tags of the other taggers).</p>
<p>The primary goal of this project is to engage the learners in a long term learning relationship with Oppia by providing them with their personal learning space - the 'Learner Dashboard'. The dashboard will keep track of the explorations and collections completed by learners, the explorations and collections that are currently in progress, the creators to whom the learner has subscribed to and the list of feedback threads which he/she needs to read or respond to. In addition, learners can also create their own curated list of explorations and collections which they may want to play later.</p> <p>The learner dashboard will help provide an enriching learning experience and will bridge the gap between learners and creators. It'll be at the heart of the learning process and help the learner learn faster. In addition, this project will create the foundation for implementing a rich set of features like recommending the right next courses for individual learners and many other such exciting possibilities. This project, I believe, would mark a major milestone in accomplishing Oppia's mission - 'to help anyone learn anything they want in an effective and enjoyable way'.</p>
<p>In recent years, “implicit 3D representations” have been attracting attention as an appropriate 3D expression for deep neural networks. At CVPR 2019, three dif- ferent solutions using this 3D representation were proposed (Occupancy Networks, DeepSDF, and IM-NET), and they show that this representation is better in some as- pects(e.g., accuracy, memory efficiency) than voxel grids or meshes, which are usually used in previous studies. Nowadays, there are two major deep learning frameworks: TensorFlow and Py- Torch. Each of them has some libraries such as TensorFlow Graphics or Kaolin, which help users to use 3D models in neural networks easily. Unfortunately, none of the three works mentioned above are implemented in TensorFlow Graphics, whereas Kaolin provides their code. This is not a good situation for TensorFlow users who want to try using "implicit 3D representations" in their experiments or projects. In this proposal, I propose a project to implement these three works in Tensorflow Graphics. I will also work on equipping some 3D geometry utilities to support users to handle the three baselines on the same interface.</p>
This project aims to automate metadata harmonization for cBioPortal, a widely used platform for cancer genomics research. Currently, metadata from different studies often uses inconsistent terms, limiting cross-study analysis and making it harder to apply AI/ML models effectively, while manual harmonization is time-consuming and not scalable. To address this, I propose building an automated metadata harmonization tool using fine-tuned biomedical sentence transformer models to semantically match and standardize metadata terms. The pipeline will include data preprocessing, semantic matching, low-confidence flagging, and expert validation, with interfaces to support expert review. The key deliverables are a fully functional metadata harmonization pipeline, semantic similarity models fine-tuned on curated biomedical data, a persistent vector database for scalable and reproducible retrieval, and documentation with testing coverage. This tool will reduce manual curation time by over 70%, improve metadata consistency, and enhance the platform’s readiness for AI-driven research.
<p>LibreHealth EHR is an open source, clinically focused EHR which was designed for easy use and customizable for variety of settings. LibreHealth is a community that works to develop and implement Health Information Technology application to deliver better health outcomes using these technology platforms one among them being the LibreHealth HER. This is an open source, clinically focused EHR which was designed for easy use and customizable for variety of settings. The back end of this EHR is forked from OpenEMR data model. To move away from OpenEMR data storage, efficient reading and writing clinical data and improving interoperability, we need to move to the standard of FHIR (Fast Healthcare Interoperability Resources) Specification. This project aims to build components for LibreHealth EHR based on FHIR with Polymer components. This will enable the EHR developers to upgrade to a standard data model. The resources from modules in level3 and level4 of FHIR with maturity level above 3 will be developed using Polymer components. These resources were specifically selected as they contribute to the patient centric feature of the EHR.</p>
<p>PyNN is a project written in python that aims at interfacing a handful of neural network simulators. A unified high level interface means that a model can be configured easily and only once and then it can be deployed on all supported simulators. Different simulators provide diverse features and also might use various numerical approaches and approximations leading to possibly different results. Having an option to compare data from several simulators without extra work is a great feature PyNN offers.</p> <p>GeNN is an efficient neural network simulator written in C++. The characteristic feature of this simulator is the ability to run simulation on GPU thus greatly decreasing required computational time.</p> <p>Interfacing GeNN from PyNN would be a valuable acquisition for both projects. PyNN will have another, faster simulator at hand and GeNN will get a python interface. Python is much more widespread among the neuroscientists. Having such an interface will help to make GeNN more popular and hopefully will reduce the time scientists spent on their simulations.</p>
The increasing adoption of AI/ML applications has led to new I/O access patterns and bottlenecks that differ from traditional scientific applications. Drishti, an interactive web-based framework for visualizing I/O traces and highlighting bottlenecks, needs to be extended to support these AI/ML workloads effectively. This project aims to make I/O behavior more accessible and interpretable for AI/ML applications by modularizing Drishti's codebase for improved maintainability, building a layer to process multiple trace files per job (common in AI/ML), and providing AI/ML application-specific recommendations to optimize I/O behavior tailored to popular ML frameworks. Deliverables include a modularized codebase, functionality to process multiple traces, AI/ML-specific I/O optimization recommendations, updated documentation for AI/ML use cases, and documentation on challenges faced and solutions when using Drishti with AI/ML applications. Ultimately, this will help AI application developers better understand and enhance their applications' I/O performance.
<p>To provide a rich experience to Octave users, the command line suggestion feature, aims to be a fast and accurate method to suggest corrections to the users whenever they make a typographic error while working on the Octave command window. It should also be able to provide the users, an easy method to select and execute the corrections suggested and enable/disable the feature. This could get tricky as Octave comes in both a GUI and a CLI. Therefore, for a successful integration of this feature with Octave, understanding how its parser, error handling, and internal functions work would be essential. The next part of the problem would be to generate the words to suggest. A trivial solution of this problem is easy but computationally expensive and hence not feasible. The challenging part therefore, is to minimize the speed-accuracy trade-off of the feature. A starting point for that would be a smart (probability based for example) implementation of the Edit-distance algorithm, using dynamic programming. Later, a Neural Network based idea would also be tested. The goal of-course, is to use the algorithm that best minimizes the speed-accuracy trade-off and provides the best UX.</p>
This project improves CloudMonkey CLI autocompletion by addressing a key limitation where enum-type parameters (such as protocol, state, and intervaltype) do not provide any value suggestions due to the absence of corresponding list APIs. While existing completion works well for commands, parameter names, and boolean values, enum parameters currently result in no suggestions, reducing usability. To solve this, the project introduces a structured allowedValues field in CloudStack’s API discovery layer, making valid parameter values machine-readable and accessible via listApis. CloudMonkey will be enhanced to consume this metadata and provide direct value suggestions, along with improvements to entity-reference completion using existing related API mappings. The deliverables include extending the CloudStack API framework, implementing enum-aware autocompletion in CloudMonkey, annotating high-impact parameters, improving entity completion reliability, and adding supporting tests and documentation. This ensures a backward-compatible, user-friendly enhancement to CLI interaction without changing existing APIs.
The objective of this project is to improve the interoperability of different blockchain networks by developing smart contract gateways that can interact with multiple networks. These gateways will automate, trace, and secure the process of submitting transactions to other blockchain networks. To achieve this goal, the project will involve creating smart contract gateways for various blockchain networks, including those using the same platform, as well as gateways to submit transactions to heterogeneous networks. The initial focus will be on developing gateway smart contracts for connecting multiple separate Hyperledger Fabric networks, and the next step will be to create gateways to connect Hyperledger Fabric to other heterogeneous networks, such as Hyperledger Indy and Besu. The project's primary objective is to evaluate system-level and large-scale integrations, which will be accomplished by setting up an extensive network of blockchains using Hyperledger Fabric as the source blockchain and other networks as destination networks. Overall, the project aims to enhance the efficiency, security, and traceability of cross-blockchain transaction submissions, thereby improving the overall functionality of blockchain technology.
Sampling from high-dimensional convex polytopes is a fundamental problem with applications in statistics, machine learning, optimization, and systems biology. Modern libraries such as VolEsti provide a rich collection of geometric random walk algorithms; however, there is currently no unified, extensible, and reproducible framework for benchmarking these methods in practice. In my recent MSc thesis, I conducted a large-scale empirical comparison of state-of-the art sampling algorithms across multiple software ecosystems. This work revealed a clear gap: while many algorithms exist, there is no standardized way to evaluate them under consistent configurations, geometries, and diagnostics. The goal of this project is to design and implement a complete benchmarking suite for VolEsti, enabling systematic, reproducible, and extensible evaluation of geometric random walks. The framework will support configurable experiments, advanced diagnostics, and integration with Python tools for visualization and analysis. Ultimately, we aim to implement a research-grade benchmarking platform where any algorithm can be evaluated quickly and thoroughly and every new suggested algorithm can be easily compared to existing methods.