NumFOCUS logo
GSoCScience and medicineAccepted in 2026

NumFOCUS

NumFOCUS promotes open source scientific software.

303
Projects
11
Years accepted
27
In 2026

Projects by year

303

2026

27 acceptedShow
  • Numba Optimized Implementations for scikit bio Performance Critical Functions

    Accepted project · by AbidKhan

    scikit-bio relies on Cython extensions for performance-critical distance and ordination operations, but these cannot accept GPU arrays. This project...

  • Lumen + Xarray Integration

    Accepted project · by AmanKumar11

    Lumen handles tabular data through DuckDB but cannot work with N-dimensional scientific formats like NetCDF, Zarr, and HDF5. This project brings...

  • Serialization Of Map and Model Classes into ASDF format

    Accepted project · by Basmala Hekal

    Gammapy currently serializes its data products across multiple data formats, which creates a fragmented workflow with poor metadata handling. This...

  • pvlib: Standardizing Diffuse Irradiance Components and Extending ModelChain's Optical Loss Modeling

    Accepted project · by Carolina Crespo

    This project aims to improve the handling of transposed diffuse irradiance and optical loss modeling in the pvlib-python package. Currently, pvlib’s...

  • Large Scale HEOM Solver Using PETSc

    Accepted project · by Chinmay Tangal

    This project aims to scale QuTiP's HEOM solver beyond single-node limitations by integrating PETSc to distribute ODE integration across multiple MPI...

  • Predictively Oriented Posteriors in PyMC

    Accepted project · by eclipse

    Standard Bayesian posteriors degrade under model misspecification by collapsing to point estimates. This project implements Predictively Oriented...

  • Native Julia ODE, SDE, DAE, DDE, and (S)PDE Solvers

    Accepted project · by HARSH_K

    OrdinaryDiffEq.jl contains over 18 IMEX Runge–Kutta implementations that share identical stepping logic but have bespoke per-method perform_step!...

  • Reduced Order Modelling with Neural Operators in Gridap.jl

    Accepted project · by Isaia Zollo

    THE PROBLEM: Solving partial differential equations (PDEs) for parametrical studies is a fundamental part of scientific research, but traditional...

  • Distributed Linear Solvers in LinearSolve.jl

    Accepted project · by Jash Ambaliya

    This project will add distributed linear solver support to LinearSolve.jl by expanding and hardening PETSc.jl integration. The goal is to let SciML...

  • AiiDA Copilot: AiiDA Natural Language Interface Using Multi-Agent AI

    Accepted project · by Jaweria Batool

    AiiDA workflows currently require manual Python scripting. This project builds a multi-agent AI system that lets scientists interact with AiiDA...

  • Stateless API for Scalable and Composable Inference in sbi

    Accepted project · by Jocho-Smith

    This project redesigns sbi's potential function API to eliminate stateful behavior, inconsistent IID handling, and device management bugs. The new...

  • Parallelizing r.proj and Raster Processing Modules in GRASS

    Accepted project · by Kaushik Raja

    r.proj and r.param.scale are widely used GRASS raster modules that run entirely on a single core, leaving modern multi-core hardware idle. This...

  • Supporting Input Convex Neural Networks in MathOptAI.jl

    Accepted project · by kofgokhan

    MathOptAI.jl is a Julia package for embedding machine learning models into optimization problems built using JuMP. These embeddings work as...

  • QuTiP: GPU-enabled circuit simulation in qutip-qip

    Accepted project · by Mudit Maheshwari

    The current gate-level circuit simulator in qutip-qip utilizes a full sparse Qobj representation of each gate, expanding local gates to the Hilbert...

  • DevOps improvement and standardization

    Accepted project · by Piyush Panwar

    This project focuses on improving the development and testing infrastructure across the ArviZ ecosystem (arviz-base, arviz-stats, and arviz-plots)....

  • Add Spatio-Temporal Datasets Support to GRASS Data Catalog GUI

    Accepted project · by Saket Kumar Mall

    Currently, the GRASS Data Catalog only queries standard spatial maps and bypasses the temporal database and STDS's entirely. As a result, Space Time...

  • DeepForest P_2: Recovering Historical Image Data Using Automated OrthoRegistration and ImageMatching

    Accepted project · by Saqlain01

    The U.S. Forest Service's Aerial Detection Survey (ADS) program has produced tens of thousands of hand-drawn polygons marking forest health damage...

  • Neural Network Builder API Refactor

    Accepted project · by Satwik Sai Prakash Sahoo

    The sbi (Simulation-Based Inference) library currently constructs neural density estimators via opaque factory functions that accept loose keyword...

  • PySal: Implementation of Geographically Weighted Matrix Decomposition Algorithms in gwlearn

    Accepted project · by ShubhamSingh1

    Standard PCA assumes a globally stationary covariance structure, missing spatial heterogeneity in geographic data. This project implements...

  • Panel / HoloViews MCP Integration

    Accepted project · by Sumaiya Islam

    Panel and HoloViews are powerful Python visualization libraries but they have no MCP integration today. AI agents can write visualization code but...

  • Trainite: A Reusable Toolbox for Training Language Models

    Accepted project · by Taha Zahid

    Every time a researcher wants to train a language model, they end up rewriting the same infrastructure: data loading, optimizer setup, training...

  • Succinct Data Structure for Efficient Operations on Trees in Scikit-bio

    Accepted project · by Timothy Wong

    Scikit-bio offers a great toolkit for working with biological data that has a tree-like structure. Although the current implementation, TreeNode, can...

  • SciML Comprehensive Non Uniform Grid Support for MethodOfLines

    Accepted project · by Utku Yilmaz

    This project aims to deliver comprehensive, end-to-end non-uniform grid support for MethodOfLines.jl. Currently, basic advection schemes lack...

  • DeepForest: Recovering computer vision annotations from historical airborne imagery monitoring

    Accepted project · by vickysharma

    After the 2010 Deepwater Horizon oil spill, airborne surveys produced 18,304 images with 2.6 million birds individually marked using a point-counting...

  • Matplotlib Overlay Layer API for Smooth Interactive Tools

    Accepted project · by Vikash Kumar_23

    I propose to implement the Overlay Layer API for Matplotlib’s interactive backends. Currently, tools like the Cursor widget force a full figure...

  • Parallelizing loo

    Accepted project · by Visruth

    The `loo` package is a widely used R package in the Stan ecosystem which exposes a clean API to cross-validate (CV) Bayesian models. Currently, some...

  • PyMC Streaming Variational Inference for Large Datasets

    Accepted project · by Yicheng Yang

    PyMC's variational inference stack — including ADVI and Pathfinder — assumes all data fits in memory. Real-world applications such as financial tick...

2025

31 acceptedShow
  • From "Stringly Typed" to "Strongly Typed" Arguments in sbi

    Accepted project · by Abel Abate

    sbi is an open-source Python package used for simulation-based inference. It provides an interface for running algorithms that are used to identify...

  • Adding embarrassingly parallel graph algorithms in nx-parallel

    Accepted project · by Akshita Sure

    Expanded and enhanced nx-parallel, a parallel backend for NetworkX that leverages joblib to parallelize graph algorithms. I implemented several...

  • State-Space Models: Dynamic Factor Models

    Accepted project · by Andrea Catelli

    State-space models (SSMs) provide a flexible framework for modeling dynamic systems where latent states evolve over time. In econometrics, Dynamic...

  • Versioning System for DISCOVER Cookbook

    Accepted project · by Ashish Mahajan

    This project focuses on implementing a versioning system for the DISCOVER Cookbook, enabling seamless navigation across distinct documentation...

  • Multidimensional SDE PINN solvers

    Accepted project · by Astitva Aggarwal

    Stochastic dynamical systems play an important role in scientific and engineering domains, particularly in systems governed by Brownian motion or...

  • SubOrganization: Aeon Project #3: Forecasting – Deep Learning for Forecasting

    Accepted project · by balgopal

    Time series forecasting is paramount in many domains, including finance, healthcare, energy, and climate science. This project suggests incorporating...

  • Community detection guide with igraph

    Accepted project · by Bea Márton

    Community detection is one of the most commonly used network analysis techniques. It is the process of grouping nodes of a network together based on...

  • Creating New Plots with Bayesplot (Stan)

    Accepted project · by Behram

    The bayesplot package is a key tool in the Stan ecosystem for visualizing Bayesian analysis results. However, it currently lacks specialized...

  • Adding Translation System for DISCOVER Cookbook

    Accepted project · by Dia29

    Currently, the Cookbook has multiple resources that can help any event became inclusive and accessible to everyone. However, its accessibility...

  • LiftOver2D: A Tool for Converting Hi-C Matrices Across Assemblies and Binning Schemes

    Accepted project · by Egor Pitikov

    The LiftOver2D project aims to develop a high-performance tool for converting Hi-C matrices between different genome assemblies and binning schemes....

  • [PYSAL] - Produce Routing Engine for `spopt`

    Accepted project · by fiendskrah

    In the Python Spatial Analysis Library (pysal), the spopt package is concerned with the formulation and solving of various spatial optimization...

  • Adding More Optimizer Interfaces to optimagic

    Accepted project · by gauravmanmode

    optimagic is a optimization library which aims to provide a unified interface to optimizers from SciPy, NlOpt and many other widely used packages....

  • Benchmarking and Profiling in toqito

    Accepted project · by justuy

    This project aims to test the performance of |toqito⟩ and compare it with other quantum computing libraries. It is divided into three parts: (1)...

  • Extending ONNX-mlpack Translator with Advanced Model Support

    Accepted project · by kumar utkarsh

    This project aims to extend the capabilities of the ONNX-to-mlpack converter by adding support for more advanced and diverse model architectures. The...

  • Adding More Optimizer Interfaces to optimagic

    Accepted project · by madman

    This project integrates optimizers from bayesian-optimization, PyGAD, and PySwarms into optimagic, a Python package that offers a unified interface...

  • Implementing the Simformer for sbi

    Accepted project · by Massimiliano Carli

    sbi is a Python package providing simulation-based inference tools for researchers and practitioners; given simulation results, sbi aims to infer the...

  • Implementing RUBI (Rule-based Integrator) in the Julia package Symbolics.jl

    Accepted project · by Mattia Micheletta Merlin

    This project aims to implement the RUBI integration system - a powerful, open source, rule-based integration system developed for the computational...

  • PyBaMM - Adding a Dispatching Mechanism for Third Party Models.

    Accepted project · by Medha-Bhardwaj

    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...

  • Implementing Integrated Nested Laplace Approximations (INLA)

    Accepted project · by Michal Novomestsky

    A key component of Bayesian inference is integrating over prior distributions to obtain posteriors. In practice however, these distributions are...

  • Create and Add to an Example Gallery for toqito

    Accepted project · by Nafisa Shamim Rafa

    This project addresses the need for enhanced usability and learning resources within the toqito quantum information library by creating a...

  • Developing an Active Learning Module for DeepForest

    Accepted project · by Nakshatra Piplad

    This project aims to enhance DeepForest by integrating and extending the active learning capabilities from Weecology/BOEM into its training pipeline....

  • Finite Volume Method Implementation for TNL

    Accepted project · by Petr Sokol

    Template Numerical Library (TNL) already includes several solvers for partial differential equations, such as the Finite Element Method (FEM),...

  • Add support for 3D thermal simulations in PyBaMM

    Accepted project · by Rishab Kumar Jha

    Seeing the current increasing demand of high performance batteries, accurate thermal modelling of battery behaviour is essential. This project...

  • Data Retriever: Creating a multi-sensor airborne benchmark for Tree Species classification

    Accepted project · by Ritesh Chowdhry

    This project aims to create a benchmark dataset for tree species classification using multi-sensor airborne data (RGB, HSI, and LiDAR) from the...

  • Integrating DeepForest Object Detection with LLM for Automated Ecological Image Analysis

    Accepted project · by Samia Haque Tisha

    This project attempts to integrate DeepForest's object detection models into Large Language Models (LLMs) so that ecological images can be queried...

  • ArviZ Plots Feature Parity

    Accepted project · by Shreyansh868

    ArviZ is a Python library for the exploratory analysis of Bayesian inference models. It is currently undergoing a major refactor into a modular...

  • aeon - Implementing and Evaluating Machine Learning Forecasters

    Accepted project · by Tina0228

    This project will investigate algorithms for forecasting based on traditional machine learning (tree based) and time series machine learning...

  • Hamiltonian Library for QuTiP

    Accepted project · by VanshajB

    This project aims to develop a comprehensive Hamiltonian Library for QuTiP (Quantum Toolbox in Python). Currently, QuTiP users must frequently...

  • Extending the loo Package

    Accepted project · by Visruth

    The `loo` package is a widely used R package in the Stan ecosystem which exposes a clean API to cross-validate (CV) Bayesian models. As new CV...

  • igraph | Degree-Constrained Null Models & Graph Construction

    Accepted project · by Zara Zong

    Graphs are powerful tools for understanding relationships and structures — from social media connections to brain networks. Often, we want to study...

  • Proposal for PYSAL– SPOPT

    Accepted project · by ZixinFeng

    This project aims to extend PySAL's spatial optimisation library (spopt) by implementing flow-based facility location models, specifically the Flow...

2024

35 acceptedShow

2023

24 acceptedShow

2022

37 acceptedShow

2021

45 acceptedShow

2020

28 acceptedShow

2019

19 acceptedShow

2018

40 acceptedShow

2017

10 acceptedShow

2016

7 acceptedShow

Public proposals

6

Read them for structure, not to copy. All proposals for NumFOCUS.