2026 Code for GoodTech (C4GT) DMP2026
Mifos has confirmed it will be participating in DMP 2026 with a set of exciting new projects.
This year we are entering 8 projects:
Mobile:
2026 Code for GoodTech (C4GT) DMP2026
2026 Code for GoodTech (C4GT) DMP2026
2026 Code for GoodTech (C4GT) DMP2026
Payment Hub EE:
2026 Code for GoodTech (C4GT) DMP2026
Mifos X - AI:
2026 Code for GoodTech (C4GT) DMP2026
2026 Code for GoodTech (C4GT) DMP2026
Mifos Gazelle:
2026 Code for GoodTech (C4GT) DMP2026
2026 Code for GoodTech (C4GT) DMP2026
Applications must be via the C4GT application Process.
Project Details:
Mobile Security Remediation 2026
Jira Top Level Ticket:
GitHub Top Level Ticket for C4GT:https://github.com/openMF/mifos-pay/issues/1998
Description:
This project is to address 13 identified vulnerabilities before new feature development.
This impacts 3 repo’s in scope:
Mifos Mobile (MM) - Epic MM-573
Mifos X Field Office App (MIFOSAC) - Epic MIFOSAC-731
Mifos Pay (MW) - Epic MW-363
It will also consider other security considerations such as ensuring the self-service plugin is used across our applications.
Mentors: Rajan Maurya
Key Skills: Mobile, KMP, Security
Application System: C4GT Application System
MVP Mifos Save Mobile Application
Jira Top Level Ticket:
GitHub Top Level Ticket for C4GT: https://github.com/openMF/mifos-save-mobile-app/issues/2
Description:
The goal of this project is to implement comprehensive Group Banking and VSLA (Village Savings and Loan Association) functionality in the Mifos Android Client application. This will enable field officers to efficiently manage group-based financial services including Centers, Groups, Joint Liability Group (JLG) loans, Group Loans with Individual Monitoring (GLIM), Group Savings with Individual Monitoring (GSIM), Collection Sheets, and meeting-based operations—all with robust offline support for remote field operations.
Key Objectives:
Center Management: Implement complete center lifecycle management including creation, activation, viewing center details, managing associated groups, scheduling meetings, and tracking center-level accounts and attendance.
Group Management: Build comprehensive group functionality supporting:
Group creation, activation, and management
Member enrollment and management
Group meetings and attendance tracking
Group-level savings and loan accounts
Joint Liability Group (JLG) operations
VSLA (Village Savings and Loan Association) Support: Implement VSLA-specific features:
Share purchases and tracking
Social fund contributions
Rotating savings mechanisms
Group fund distribution (share-out)
Member savings and loan tracking within VSLA cycle
Group Loans with Individual Monitoring (GLIM): Enable field officers to:
Create group loans with individual member amounts
Track individual disbursements within group loan
Monitor individual repayment schedules
Process individual and bulk repayments
View individual vs group loan performance
Group Savings with Individual Monitoring (GSIM): Support group savings accounts where:
Individual member contributions are tracked
Mandatory savings linked to loan repayments
Individual balances within group account visible
Bulk deposits and withdrawals supported
Collection Sheet & Bulk Operations: Implement efficient collection tools:
Productive Collection Sheet (staff-based)
Center/Group Collection Sheet
Bulk loan repayments
Bulk savings deposits (mandatory savings)
Attendance tracking
Meeting-based transaction processing
Offline-First Architecture: All group banking operations available offline with automatic sync when connectivity returns.
Mentors: Rajan Maurya
Key Skills: Mobile, KMP,
Application System: C4GT Application System
Mobile [Flow] Pocket Linked Accounts
Jira Top Level Ticket:
GitHub Top Level Ticket for C4GT:https://github.com/openMF/mifos-pay/issues/1997
Description:
This project implements the “Pocket” feature - enabling users to organise their financial accounts (savings, loans, shares) into a unified pocket view. This is both Internal and External accounts so needs to consider PISP.
This is part of Mifos Pay (MW) - EPIC MW-378.
Mentors: Rajan Maurya
Key Skills: Mobile, KMP
Application System: C4GT Application System
Payment Hub EE, Operations UI in React
Jira Top Level Ticket:
GitHub Top Level Ticket for C4GT: https://github.com/openMF/ph-ee-operations-web-react/issues/2
Description:
This project will create a new operations application from the ground up for the Payment Hub EE solution target at release 2.1 (Oct 26).
This task will need to replicate as an MVP the functional tasks of the existing Operations App (Angular)
Support both GovStack and non-GovStack use
Orchestration for Governments
G2P
P2G
Vouchers
Orchestration for DFSPs
Integrate with RBAC (keycloak)
different views for different roles
authenticated login
Integrate with Reporting/Dashboards.
Easier view for bulk transactions
failure rates
etc
Steps:
Research best practice Design in Figma the ideal screens
Identify MVP scope
Modularise components within the UI so only those needed are deployed.
Plan for extensibility how would new functions be added for new PH-EE capabilities
Plan for maintainability how will the app be maintained
Test Coverage from start building suitable test cases for regression testing
Mentors: David Higgins, Victor Romero
Key Skills: React ShadCN
Application System: C4GT Application System
Agentic AI for Proactive Portfolio Management with Explainable Autonomy
Jira Top Level Ticket:
GitHub Top Level Ticket for C4GT:https://github.com/openMF/mifos-x-ai-agentic-framework/issues/2
Description:
Overview and Objectives: Design and build a prototype "Portfolio Health Agent" that autonomously monitors loan portfolios, identifies at-risk accounts, and takes pre-approved actions (e.g., sending reminders, scheduling follow-ups) while escalating complex cases to human loan officers. The project explores how to balance AI autonomy with explainability and human oversight in a regulated financial context.
Description (Glance): Develop an agentic AI framework using LangChain and a local LLM to proactively manage loan portfolios. The agent fetches data from Mifos X, applies risk scoring, and executes configurable actions with human-readable explanations. A dashboard in the Mifos X Web App allows loan officers to review agent decisions, provide feedback, and adjust autonomy levels. The project includes a user study to evaluate trust and usability.
Resources:
LangChain for agent building:
https://www.langchain.com/FINRA guidance on agentic AI:
https://www.investmentnews.com/regulation-legal-compliance/finra-flags-rise-of-agentic-ai-seeks-member-firms-feedback/264996Anthropic's research on interpretability:
https://www.anthropic.com/researchMifos X API:
https://demo.mifos.io/api-docs/apiLive.htmMifos Repositories:
openMF/mifos-x,openMF/web-app,New repo for solution:
openMF/mifos-x-ai-agentic-framework
Impact: Creates a reusable framework for explainable financial AI agents, improves portfolio quality, and demonstrates responsible AI autonomy.
Channels to Discuss: Slack: #wg-ai-for-all, #mifos-x-dev
Mentors: @Priyanshu Tiwari, @Akshat Sharma, @Rahul Goel
Key Skills: Python, LangChain, Prompt Engineering, Agentic AI concepts, Explainable AI, Basic frontend (React/Angular), Research methodology.
Application System: C4GT Application System
Dynamic Pricing of Micro-loans Using Reinforcement Learning
Jira Top Level Ticket:
GitHub Top Level Ticket for C4GT: https://github.com/openMF/mifos-x-ai-sentiment-analysis-module/issues/2
Description:
Overview and Objectives: Build a simulated environment to test dynamic pricing strategies for micro-loans using Reinforcement Learning (RL). The project explores how RL agents can learn personalized interest rate policies that balance institutional sustainability and client welfare, while incorporating fairness constraints and safe exploration.
Description (Glance): Design a Gymnasium environment modeling client characteristics, loan offers, and repayment dynamics calibrated with historical Mifos X data. Train RL agents (e.g., DQN, PPO) using multi-objective rewards. Implement fairness constraints and safety layers, evaluating agents for disparate impact and robustness to economic shocks. Document findings and trade-offs.
Resources:
Hugging Face multilingual models:
https://huggingface.co/models?pipeline_tag=text-classification&language=multilingualXLM-RoBERTa paper:
https://arxiv.org/abs/1911.02116UNESCO's work on language preservation and AI:
https://en.unesco.org/artificial-intelligence/language-preservationLabelStudio for annotation:
https://labelstud.io/• Mifos Repositories:
https://github.com/openMF/mifos-xhttps://github.com/openMF/web-appNew repository:
mifos-x-ai-sentiment-analysis-module
Impact & Research Contribution:
Academic: Contributes novel research on adapting NLP for financial sentiment analysis in low-resource languages and culturally diverse contexts, with a focus on predictive validity rather than just classification accuracy.
Practical: Provides MFIs with a tool to monitor client well-being at scale, enabling proactive interventions that can prevent defaults and support clients during hardship.
Social: Moves beyond purely financial metrics to incorporate human-centered indicators of client health, aligning with the mission of financial inclusion as a tool for human development.
Channels to Discuss: Slack: #wg-ai-for-all, #mifos-x-dev
Mentors: @Priyanshu Tiwari, @Akshat Sharma
Key Skills: Python, PyTorch/TensorFlow, Gymnasium, RL algorithms (Stable-Baselines3), Fairness in ML, Research methodology.
Application System: C4GT Application System
Incorporation of Mifos X latest version into Mifos Gazelle
Jira Top Level Ticket:
GitHub Top Level Ticket for C4GT: https://github.com/openMF/mifos-gazelle/issues/102
Description:
Mifos Gazelle currently doesn't deploy the latest version of Mifos X this project is to update the deployed version of Mifos X to the latest version.
This includes ensuring the deployment includes all elements such as reporting modules, credit bureau analysis modules etc.
The project must also identify and document a way this can be undertaken repeatedly in the future as new versions are released with minimal effort required. This includes test cases for regression. How extensibility would be handled.
Memory utilisation and optimisation including using shared resources is key for Mifos Gazelle and this will need to be considered as part of this project.
As a first step interested applicants should look and investigate the current Mifos X deployment in Mifos Gazelle
Mentors: Tom Daly, David Higgins, Victor Romero?
Key Skills: Helm, DevOps, Python, JS, Kubernetes
Application System: C4GT Application System
Mifos Gazelle JVM heap space project with configurable environment views
Jira Top Level Ticket:
GitHub Top Level Ticket for C4GT:https://github.com/openMF/mifos-gazelle/issues/103
Description:
The idea behind this project is to optimise JVM heap space for Mifos Gazelle for different environments using profiles.
Environments that should be supported:
Raspberry Pi running Ubuntu. (micro environment)
VM running Ubuntu. (std environment)
Multi VM Kubernetes cluster for performance transactions (perf environment)
In some ways this is similar to the concept that used to exist in PH-EE of different configurations
This project will have support from our Mifos Gazelle Mentoring team, but will ideally suit someone who with guidance is willing to investigate and come up with solutions.
Mentors: Tom Daly
Key Skills: JVM, Kubernetes, DevOps, Helm
Application System: C4GT Application System