Google Summer of Code 2024 Ideas

The Mifos Initiative was accepted as a mentoring org for the 2024 Google Summer of Code. We are continuing to update and refine our list of ideas and assigned mentors. Attend our AMA on March 7 at 1500GMT. Applications open up on March 18. 

Get Stoked - End Poverty. One Line of Code at a Time.

 



2024 Google Summer of Code - Get to Know Mifos

We're looking forward to participating in Google Summer of Code for our twelfth year. In 2023 we worked with a total of 14 interns through Google Summer of Code and our inaugural Mifos Summer of Code across the back-end platform, our web and mobile apps, and our AI tools. We hope to continue building our next generation of contributors who are joining in our movement to fight poverty through open source software. We want you to be part of our mission of creating a world of 3 Billion Maries.  

Core DPG Technology Stack

GSOC Contributors in 2024 will have the option to work on a variety of projects related to our end to end open source stack for digital financial services including open source core banking, generative AI for financial services, real-time payment and instant payment integration with Mojaloop and Interledger Protocol, digitizing government to person (G2P) payments, mobile money and Open Banking APIs, and our suite of cutting edge and web and mobile apps. Participating interns will get to work on our new front-end Mifos X web app built on Angular, our suite of mobile apps including field officer apps and mobile banking apps for Fineract, our mobile wallet and integrations with the GSMA mobile money API and Open Banking API.  We will once again have projects related to our award-winning Payment Hub EE which provides an orchestration engine to connect to real-time payment systems like Mojaloop, ACH, and emerging payment protocol like Interledger. Given the core stack is mature in its functionality, we continue to focus GSOC projects on maintainability and deployability, security, improving developer experience, user experience optimization, and more.

Digital Public Infrastructure

With the international development’s sector deepened focused on digital public infrastructure, we are also seeking interns to contribute to our efforts with open source digital public goods for G2P payments like our Payments Building Block powered by Payment Hub EE and OpenG2P, the initiative we've helped to launch to digitize large scale government cash transfer programs and parallel projects for in-kind transfer and management of social registries like OpenSPP. This year, we’ll also have a heightened focus on AI as part of AI for All and Data Science for Good initiatives.

AI for All

AI and Machine Learning will have a profound impact on financial servics. It’s a key focus area of our community and we have a variety of projects that build upon our existing AI tools as well as exploratory ones to discover and innovate around the power Generative AI and LLMs and LAMs can have on digital financials services for the Base of the Pyramid. A number of ideas will focus on continuing to refine and enhance our existing AI projects including a chatbot framework for improved customer support, credit scoring module using machine-learning, and our vision PPI app using Cloud Vision API. Exploratory efforts underway in the community that build upon existing leadership and expertise of our mentors an need across the sector include AI for agricultural credit scoring, AI for bank statement analysis for credit decisioning, generative AI for streamlining how to configure and extend Fineract, using NLP to add additional language support to our chatbot, LAMs to enable intiatition of transaction flows via our banking apps, POCs to explore the usage of tools CoPilot to improve our developer experience and ease of extending Mifos, Generative AI to improve end user documentation, and AI for fraud detection and transaction monitoring.

Modularization on the Back and Front-End

We have some major refactoring and modularization of Fineract 1.x that contributors could continue to help with. As we modularize the back-end we are also moving towards modular resuable UI components across our web and mobile apps exploring the usage of new frameworks like Compose Multi-platform. Through re-usable UI components, cross-platform development, frameworks, and improving our mobile SDKs, we are seeking to significantly streamline the development and design of our mobile apps.

Integrations & POCs

Since the boundaries of fintech and financial services extend so far beyond the core we also have a number of interesting POC integration projects lined up by our mentors this year including POCs with alternative reporting engines, mobile check deposit POCs, workflow engine integration, integration with KYC automation frameworks, POCs for selfie verification, and more. We also will continue to deepen our integration with peer projects in the DPG and Financial Inclusion space such as Mojaloop for Instant Inclusive Payments, MOSIP for Digital Identity and eKuta for Fraud and Risk Management.

Payments

While both the back and front-end development of our core banking DPGs will be a major focus, we will have more projects this year on top of our Payment Hub EE which provides an orchestration engine to enable the ease and participation of fintechs and financial institutions into modern payment systems. This year we will continue to improve upon the user experience of our operations app and control center, make it easier to build mobile money connectors, deepen our integration with instant inclusive payment systems like Mojaloop, test out integrations with new solutions like Interledger Protocol and Rafiki, leverage open payment and banking standards.

Small Projects

With the new category of smaller projects, we do have a number of projects that are more research and POC-based including:

  • Alignment with emerging Open Wallet Standards

  • Integration with Open Banking and Open Payment Standards

  • Design and Adoption of new frameworks like Compose Multi-platform

 

Mifos is a globally recognized digital public good and an award-winning project that sits at the cutting edge of inclusive fintech and embedded finance. Mifos is the leader of the open source banking movement helping to commoditize core banking infrastructure enabling financial services to be embedded anywhere and everywhere.

The best way to understand what we do is to watch a few videos. 

What is Mifos and What Does our Community Do?

 

https://youtu.be/pE28NrY66Zc

Why Does Google Summer of Code matter so much to Mifos and what do we look for in contributors?

How is the Mifos software used? 

  

Guidelines

  • Getting started Read about setting up the code and understand the basic concepts around MifosX.

Expectations

Contributors working on Mifos X will be expected to:

Prerequisite Skills

Basics 

  • Be a quick learner

  • Be well-behaved, act in good faith and be of good humour.

  • Troubleshooting Wizard

  • Passion for writing beautiful code

  • Excellent communication skills

  • Knowledge of developer tools

    • such as a text editor, source control, how to build software

    • experience with specific tools will also help, such as Eclipse IDE, Git, IntelliJ

Mifos includes a wide variety of technologies, we do not expect a student to be an expert on all of these. But it will be helpful if you have some experience in some of these. Helpful skills (specific technology requirements vary with project chosen) and must be eager to learn and develop with the requirements:

  • Java, Spring, MySQL, Jersey & Hibernate

  • HTML, CSS, JavaScript (JQuery), Angular & Material Design

  • JUnit, REST-assured

  • Android

Source Code

Mobile App Overview: https://openmf.github.io/mobileapps.github.io/

Mifos X and Apache Fineract 1.0 (Gen 2)

Hints

  • When you need help, ask for help after exploring all options on the web. We are very excited for you to join us, but we need to know that you're willing to put in the time and effort required to do your part. When you do ask, ask well.

  • Tips for a Good Application from former GSOC intern and Mentor, Ishan Khanna: https://hackernoon.com/7-things-you-need-to-know-to-ace-your-gsoc-proposal-8e422f2b6abe

  • Not sure if you are qualified? Download and build the code, then run the Platform and the Mifos X distribution.

  • Email the mifos-developer mailing list with your questions and project ideas.

  • Install Skype for (generally infrequent) video calls. You'll need a headset for Skype, too.

  • Get started by solving few issues on Fineract here or for the web app here.

Project Ideas

In Progress

Our 2024 Ideas list is currently being refined. Stay tuned for additional updates. 

2024 projects will be related to the Mifos X Web App, our mobile apps including our Android Field Operations app, our Mobile Banking app, or our Mobile Wallet App or our Online Banking App. There will also be projects focused on our Payment Hub EE payment orchestration engine and our efforts in Digital Public Infrastructure space around G2P Payments for Social Protection Some projects will also center directly around the Apache Fineract platform for financial inclusion and additional tools, features, or modules and integrations with solutions that sit around the core platform. All the Mifos apps and modules are built on top of the Apache Fineract platform, which Mifos developed and donated to the Apache Software Foundation. Fineract is the world's only open platform for financial inclusion and is banking delivered as a service via the cloud. Apache Fineract is a true platform in which the back-end is cleanly separated from the front-end and all core platform services are exposed through an API making it easy to develop new applications on top of the Apache Fineract platform. 

Note: The Apache Fineract project is not managed by Mifos. The Apache Fineract PMC (project management committee) has recently made decisions to strongly focus on stability, quality, and maintainability of the platform. While the project will continue to welcome Pull Requests (PRs) from contributors, there is now a stronger expectation of independent quality assurance steps done prior to any code being accepted into a release branch. The fineract project is discussing the idea of an “experimental branch” as a way to allow for GSOC volunteer improvement to be have their work shown. GSOC volunteers should expect more scrutiny for any code submission, including full unit tests, regression tests, performance tests, and maintainability of the solution.

Large Projects

For 2024, GSOC projects can be of three durations - large (350 hours) and medium (175 hours) and small (90 hours). These projects are of the 350 hour duration.

Refactor Android Client & Rewrite using Android SDK

Mentors

@Chinmay Kulkarni @Shashank Priyadarshi

Length

Large - 350 hours

Category

Mobile - Android SDK | Core Development

Overview & Objectives

The goal of this project is to upgrade our Android field officer application to its latest dependencies and to rewrite the application to consume the recently developed Android SDK written using coroutines. The Android client will be the first of the Mifos mobile apps which consume the SDK, reducing repeated network layer code in mobile apps, improving developer experience, and making it easier to migrate to newer versions of Apache Fineract 1.x.
and as well as student would be working on migrating xml to jetpack compose to support multi-module MVVM architecture and adding new features with jetpack compose.
This year we are totally focused on migrating to Jetpack compose and support KMP (Kotlin Multi-Platform)

Description

In 2024, The student will be working on implementing the following things:

  1. Update the Android client to latest dependencies

  2. Rewrite Network layer of Android client to consume SDK using coroutines instead of RxJava implemented code.

  3. Migrate xml implemented features to jetpack compose

  4. Implement multi-module MVVM

  5. Add kotlin multi-platform support.

  6. Write appropiate unit and integration tests.

  7. Update CI/CD to build APK and analyses code quality

  8. Implemented Playstore release github action pipeline

  9. Integrate google playstore API for better and faster release flow.

  10. Update corresponding documentation for building the app.

Helpful Skills

Android, Kotlin, Jetpack compose, navigation-compose, MVVM, coroutines, Flow, multi-module architecture.

Impact

  1. High performance using jetpack compose

  2. Clean code and easy understandable code because of multi-module and coroutines implementation.

  3. Overall stability by increased testing coverage through a more stable and error-free codebase.

  4. Improved developer experience through faster build time.

  5. More seamless release management and upstream contribution

  6. Reduction of time to upgrade

Other Resources

QA & Testing - https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=195730042

Github

https://github.com/openMF/android-client

https://github.com/apache/fineract

https://github.com/openMF/fineract-client

https://github.com/openMF/fineract-android-sdk

Expand Unit Testing Coverage of Fineract with Cucumber Testing Framework

Mentors

@Aleksandar Vidakovic

Length

Large - 350 hours

Category

Platform | Core Development | Infrastructure

Overview & Objectives

In 2023 we have made substantial progress in implementing a baseline of behavior-driven tests across the platform using the Cucumber framework. The goal of this project is to expand unit testing coverage across the Fineract platform. Currently most of our automated testing is only through integration tests which take a long time to run and aren’t consistent. Cucumber is being implemented as the unit test framework and this project would focus on converting existing integration tests to unit test and writing new unit tests.

Goals are to increase testing coverage of core modules, reduce run-time at build of completing tests, and implementing some automated reporting to show testing coverage.

Description

The student will be working on implementing the following things:

  1. Collaborate with mentor to implement Cucumber framework

  2. Collaborate with mentor to implement test containers

  3. Refine test data set and scripts

  4. Convert high priority integration tests to unit tests

  5. Write unit tests for key functional modules

  6. Implement reporting to show test coverage.

Helpful Skills

Cucumber, Java, Spring

Impact

  1. More frequent & higher quality releases

  2. Improved developer experience and ease of submitting pull requests.

Other Resources

 

Github

https://github.com/apache/fineract

 

Functional Enhancements to Mobile Wallet for G2P Use Cases

Mentors

@Rajan Maurya @Avinash Vijayvargiya

Length

Large - 350 hours

Category

Mobile - Mifos X | Core Development

Overview & Objectives

2024 development will focus on refining our current Mobile Wallet to be a strong reference implementation of feature-rich and secure mobile wallet application for G2P use cases. While our mobile wallet is a powerful tool for all fintechs and financial institutions, we want to powerfully and simply demonstrate the ability for governments to disburse G2P Payments to highly functional mobile wallet and then be able to transact with that wallet by paying for bills, sending transfers, and paying for goods and services via merchants. We provide a reference mobile wallet application for consumers and merchants that has been developed by our Google Summer of Code interns from 2017 to 2023. The mobile wallet provides an extensible mobile wallet framework to support the basic use cases of a mobile wallet as documented in the Level One Project mobile wallet requirements. This extensible framework should support both merchant and client use cases as well as be capable of integrating with a Fineract back-end

Over time, we would like Mifos X to be more generically a wallet management system and this reference application is a powerful tool to support that. 

Description

The initial mobile wallet framework along with 2 reference apps, PixieCollect and MifosPay, were developed in 2017. Later we decided to continue with MifosPay application only which as mentioned uses the mobile wallet framework.

In 2019, these functionalities were extended further by Shivansh including improving user experience and redesigning the app, support for Kotlin, integration with two Mojaloop transaction flows via the Paymeht Hub, adding improving Deeplinks, support for standing instructions and more well-rounded support for merchant transactions.

In 2020, Devansh Aggarwal further added complete support for standing instructions, integrated with Fineract CN for core use cases by mapping Fineract back-office APIs to Fineract CN APIs, added multi-theme support, completed integration with Payment Hub EE, added support for Hover, and converted Java code to Kotlin (in progress). For more details refer this

In 2021, Kinar Sharma, worked on developing a new multiplatform mobile wallet application using Kotlin multi-platform. This new application consumes FineractCN APIs and is built upon clean architecture. Kinar completed the data, domain and presentation layer (only for Android) for usecases available in FineractCN.

In 2022, Prashant Singh continued to evolve the app.

In 2023, Rachit focused on refining the architecture, streamlining the design and implementing some of the G2P use cases.

In 2024, we are targeting to make production-ready our cases for G2P, update dependencies, consume more uniformly our SDKs fix pending issues and introduce new features. Functional enhancements include:

  • Integrate latest version of Payment Hub EE

  • Integrate Mifos' notifications framework to provide support for usecases like merchant request to pay.

  • Incorporate Mifos' new design library

  • Migrate xml to Jetpack compose

  • Basic implementation of multi-platform

  • Implemented Playstore release github action pipeline

  • Update wallet framework to be make use of Mifos' Android SDK

  • Complete migration of Java code to Kotlin

  • Improving the security framework to integrate more seamlessly with middlewares, API gateway, and identity management software

  • Exploring proof of concept architecture or redesigns to align with movements like the Open Wallet Foundation.

Helpful Skills

Android development, Java, Kotlin, Jetpack compose, Rest, SQL, Git

Impact

By providing an extensible mobile wallet framework, allow partners a complete reference stack of back and front-end applications to offer digital financial services to clients.

Other Resources

2023 Mobile Wallet Final Report: https://docs.google.com/document/d/1xHDVzmxuBj5KA0vHiJS_UGrs0Zb6IrpqwRmSdW3uExM/edit#heading=h.h4y8eq18ezr9

2020 Mobile Wallet Progress: https://gist.github.com/devansh-299/e2041c07d9ab55a747391951e9090df4

Mobile Wallet Framework: Source Code | Issue Tracker | Gitter Chatroom | Slack

See https://openmf.github.io/mobileapps.github.io/

Integrate Fineract Client with Workflow Engine

Mentors

@Aleksandar Vidakovic @Victor Romero

Length

Large - 350 hours

Category

Fineract Platform | Modules

Overview & Objectives

Users of Mifos and Fineract have long had a need to have greater control and flexibility over creating loan and customer onboarding workflows that incorporate internal processes/steps as well processes involving external systems. This project would center around creating an external integration with a workflow engine such as Flowable or jBPM using the Fineract REST API as the glue. The Fineract Client library enforces the contract between the workflow engine and Fineract. Fineract Client and its REST API as the glue. The result would be a UI-driven workflow engine to allow non-technical users to define these new custom workflows where they could drag and drop the different steps of the process. 

 

 

Description

Assumptions:

  • BPMN should be adopted as a standard so which can make the workflow engine independent of the technology stack. 

  • BPMN editors allow end users to sketch the workflow and generate a machine readable and executable output. 

  • Workflow actions or business logic would be written in Java as part of Fineract so adapters will be need to be created on the workflow engine to execute actions on our side. 

  • External processes would be sub-flows that could be used like lego bricks and included in the parent workflow context. 

 

Steps:

  • Intern would select a workflow engine to integrate with and then build adapters to trigger and transfer form BPMN form data into specific Fineract REST API calls. 

  • Currently business logic such as creating a customer is triggered by an API call in Fineract. For this project, the intern would define a BPMN form with all the data that is needed which could be defined as a human task. Proceeding through each step triggers an adapter that transforms the generic BPMN form data into specific Fineract REST API calls

  • Intern would go through the entire REST API building these adapters to trigger transformation of BPMN into Fineract REST API callsCurrently when creating a customer, API by API, the intern would define 

  • Once all the calls in Fineract have been transformed, the integration could provide for triggering any external  customer-specific action/workflow/API call/ system integration

Helpful Skills

Java, BPMN

Impact

UI-Driven interface based on BPMN standard to allow for non-technical users to define customer onboarding and application workflows via a drag and drop interface.

Other Resources

https://github.com/flowable

https://www.jbpm.org/

https://www.bpmn.org/

Extend and Evolve UI Library of common components across all Mobile Apps

Mentors

@Chinmay Kulkarni @Devansh Aggarwal @raul.sibaja

Length

Large - 350 hours

Category

Mobile - Mifos X | Core Development | Infrastructure

Overview & Objectives

In 2022, Rahul Gill, created and completed the first iteration of our UI library for our mobile apps. Our suite of customer-facing mobile applications include our mobile wallet framework, and mobile banking apps for Fineract 1.x and Fineract CN. These are designed to serve as reference implementations for demonstration purposes but also to act as secure and robust starting dough that can be extended and enhanced and white-labeled.

With the move towards more digital financial services, these reference solutions are ever more important and critical and must appear highly polished, clean, and professional. We are working with a designer to provide a set of clean, consistent and professional UI designs and workflows to implement across our customer-facing apps.

This project would focus on implementing these new designs across the customer-facing apps providing a consistent and familiar look and feel. It will build off of efforts in 2020 and 2021 implementing the UI designs previously proposed during GCI. 

Across all our mobile apps, there are common screens and workflows with a lot of redundant and inconsistent design and development from scratch. The creation of a UI library of common shared components and design standards and guidelines would enable the following:

  • Improve developer experience and ease of development

  • Consistent look and feel of UIs for all apps

  • Defined process for updating apps.

  • Develop common UI library to ensure consistency of all apps

  • Design guidelines, principles, and standards

  • Could potentially be a valuable upstream project in and of itself to create mobile fintech apps

Description

In 2024, With leadership of our mentors, this project would focus on building on the initial shared components of the UI library based off of the common screens and workflows identified across the various mobile apps. These common flows will be broken down into their-base-level elements and components.
If time permits, these common components could be tested and consumed by building out the UI screens for our reference open banking fintech app using this library.

Design enhancements to customer-facing apps include: 

  • Break down common workflows into base-level components

  • Refine and update design standards and guidelines

  • Create and set up repository to house elements, components, and designs using jetpack compose.

  • Design core ui component for our fintech app using jetpack compose.

  • Develop and create base-level elements and components accoring to UI design standards and guidelines

  • Implement screens and workflows to test out on reference open banking fintech app

  • Documentation to ensure how to use UI library and update mobile app when UI library is updated.

Helpful Skills

Android Development, Kotlin, Java, Jetpack Compose, XML, Git

Impact

A clean and simple UI is key for our low-tech audience and professional and consistent look and feel enhances credibility of our stack. 

Other Resources

Recap on 2022 GSOC Project from Rahul Gill: https://gist.github.com/rahul-gill/6700c366627a2c7007a07b96ff6c0785

https://mifosforge.jira.com/wiki/spaces/MIFOSX/pages/3013476363

2020 UI Enhancements: https://gist.github.com/ShivangiSingh17/67b6041387c1e281caa7df23347f549e

Mobile Wallet Framework: Source Code | Issue Tracker | Gitter Chatroom

Mifos Mobile - Android Mobile Banking App: Source Code | Issue Tracker | Gitter Chatroom

See https://openmf.github.io/mobileapps.github.io/

Implement Kubernetes Operator for Mifos/Fineract and/or PH-EE

Mentors

Tom Daly

Length

Large - 350 hours

Category

Platform | DevOps | Payment Hub EE

Overview & Objectives

The goal of this project is to improve the deployability of key projects of our stack like Mifos/Fineract and Payment Hub EE.

Kubernetes operators give kubernetes application dev/ops a great deal of reliability and architectural control over kubernetes applications deployments. Reference https://kubernetes.io/docs/concepts/extend-kubernetes/operator/ for details on the benefits of Kubernetes operators. The benefits of Kubernetes operators can also be quite tangible and visibile by looking at those under development from peer projects in the financial inclusion space like Mojaloop vNext whom are building out a Kubernetes operator as a means to improve quality, simplify deployment, and improve reliability and testing.

Description

This project would focus first on gathering the learnings around Kubernetes operators and their patterns. There are many good examples in existence to follow (i.e. Mongo, Kafka, etc). This learning around the design would compose 30% of the project. The next and major focus of the project would be to choose what to move from what’s currently only in Helm to an Operator. Likely there will still remain a small helm chart simply to deploy the operator.

Helpful Skills

Docker, Kubernetes, Jenkins, Bash, Java - Spring, PostgreSQL, MariaDB, Cassandra, TDD With JUnit 4, Gradle

Impact

DevOps and Sys Admins running our projects would experience a more simplified deployment with greater degree of control, improved quality of their builds and greater reliability and ease of testing.

Other Resources

LAM (Large Action Model) for Fineract (AI)

Mentors

@jeremy engelbrecht

Length

Large - 350 hours

Category

AI | Platform - Modules | Exploratory

Overview & Objectives

The idea behind this project is to use a LLM to give command/prompts and have those commands fulfilled via selenium

Description

Following the similar types of commands/actions that can be executed in Fineract via our chatbot, this project would use LLMs to determine the intent of a user and fulfill those actions in the Mifos/Fineract applications using a tool like Selenium.

Helpful Skills

Python, LLM(Llama2 or similar), Selenium

Impact

This would dramatically change the way individuals interact with financial services. It goes well beyond a chatbot by being able to engage with a tool that can apply for and initiate financial transactions via prompts.

Other Resources

https://xgboost.readthedocs.io/en/stable/

Generative AI to Improve Mifos Documentation

Mentors

@jeremy engelbrecht @Lalit Mohan S

Length

Large - 350 hours

Category

AI | Platform - Modules | Exploratory

Overview & Objectives

Mifos and Fineract and the suite of core banking applications we provide are highly complex applications both from a technical and domain knowledge perspective. Maintaining documentation around installing, configuring and using the application is challenging and presents a steep learning curve.

The goal of this project is to use Generative AI to create a Transformer such as LLM to have a Q&A with the organisation documentation. By having this LLM trained across the user manuals, wiki, ReadMEs, mailing list posts, forum posts, slack discussion, and documentation on our project, it would greatly simplify the experience for any implementer looking to use the software.

Description

Intern would create a Transfer such as LLM trained on all the sources of technical and user documentation across our project such that implementers could intereact with the documentation in a QA style format.

Helpful Skills

Hugging Face transformer, Llama2 or similar, Colab

Impact

This would greatly simplify the process of configuring, deploying and using our core banking software across various use cases making the software and documentation more maintainable for the project and more usable for the customer base.

Other Resources

https://huggingface.co/docs/transformers/model_doc/llama2

Fraud & Risk Management & Transaction Monitoring POC (AI)

Mentors

@jeremy engelbrecht @Lalit Mohan S @Aleksandar Vidakovic

Length

Large - 350 hours

Category

AI | Platform - Modules | Exploratory

Overview & Objectives

With continually growing adoption of Payment Hub EE to connect fiancial institutions using or not using Fineract into real-time payment systems, the likelihood of fraud is ever-growing. With all of this transactional data flowing into and out of Fineract, there is now also a wealth of data to analyze using rule-based and AI-based methods to detect transactional patterns and identify fraud. A lot of fraud detection and transaction monitoring happens centrally at the level of the payment switch but there’s quite a bit of value in analyzing this at the level of each individual DFSP and sharing this on-us transaction data back to the central payment switch or system

Goal of this project would be to work on a proof of concept integration between Mifos/Fineract and Payment Hub EE with a FRMS solutino for fraud detection and transaction monitoring. Given the power and potential of generative AI for financial services, project shoudl explore the use of generative AI for the following areas as documented by Newron:

  • Anomaly Detection: Generative AI models can create synthetic data that mimics legitimate transactions. By training on both genuine and synthetic data, models become adept at detecting unusual patterns indicative of fraud.

  • Adaptability: Unlike rule-based systems, Generative AI models can adapt to new fraud patterns. As fraudsters evolve, the model evolves with them, improving detection rates.

  • Data Augmentation: Generative AI can augment imbalanced datasets by creating synthetic examples of rare events (fraudulent transactions), making the model more robust.

  • Unsupervised Learning: Generative AI models, like Variational Autoencoders (VAEs) or Generative Adversarial Networks (GANs), learn patterns in an unsupervised manner, reducing the need for manual labeling of fraudulent cases.

  • Behavioural Analysis: Generative AI can analyze transaction behaviours, such as user interactions or purchasing habits, to identify deviations from the norm

Project could also integrated with eKuta which although rule-based and not AI-based provides a powerful FRMS solutino that aligns and integrates nicely with IIPS like Mojaloop.

Description

Intern would work on a POC integration to monitor and analyze data from real-time payment system flowing via Payment Hub EE into Fineract. Project could utilize generative AI to improve effectiveness of solutiona and can leverage existing efforts in the community to collaborate with eKuta solution.

Helpful Skills

Hugging Face transformer, Llama2 or similar, Colab

Impact

Real-time payments make the challenges and consequences of fraud ever more steep and severe. Using Generative AI to combat it and integrating with FRMS solutions will help the user base using Payment Hub EE and Fineract.

Other Resources

Open Banking/PISP Fintech App Framework Version 3.0 (GovTech)

Mentors

@Naman Dwivedi @Shivansh TiwariKarim Jindani @

Length

Large - 350 hours

Category

Mobile, Exploratory

Overview & Objectives

Across our ecosystem we're seeing more and more adoption and innovation from fintechs. A huge democratizing force across the financial services sector is the Open Banking movement providing Open Banking APIs to enable third parties to directly interact with customers of financial institutions. We have recently started providing an Open Banking API layer that will allow financial institutions using Mifos and Fineract to offer third parties access to requesting account information and initiating payments via these APIs. Most recently the Mojaloop community, led by Google, has led the development of a centralized PISP API

To demonstrate these Open Banking APIs and use cases that third parties and fintechs can provide we have developed a cross-platform reference mobile app on Kotlin to showcase a number of these features. It currently connects with the Open Bank Project that adheres to the UK Open Banking API standard. The API Gateway to connect to is still being chosen (WS02, Gravitee, etc.)

The breadth and variety of apps that could be built leveraging these APIs from region to region is endless. We would like this app to be built in an extensible and modular fashion such that core libraries and components could be re-used across different use cases with this framework as the foundation and multiple reference apps on top. Applications include personal financial management apps aggregating information from multiple bank accounts in one place, wallet apps allowing payments to be made from different banks, lending apps, leveraging data and insight from multiple accounts, savings apps, etc.

Description

Intern would work on refining the initial architecture of the framework, the UI and user experience, core use cases including customer authentication and onboarding that was implemented in 2020 and integrating with the Fineract Open Banking APIs and Mojaloop PISP APIs to demonstrate use cases around account information request and payment initiation. 

This would be a continuation of the project worked on a couple year’s back but essentially the idea is the same to have a stand-alone reference third party fintech app that demonstrates 3rd party initiation/ Open Banking APIs so the app itself would need its own separate user managment, etc but then it would pull in data by authorizing consnet via open banking api to pull in transactional data from a mifos/fineract account (could extend exiting open banking adapter and map additional fineract apis to UK open banking api standard and most importantly we’d want to use this reference fintech app to demonstrate the use of Mojaloop/Google 3PPI PISP APIs whereby a user could authorize and establish consent across any bank participating in a mojaloop switch to the fintech to initiate transactions through mojaloop via their accounts managed in Mifos/Fineract

  • Aggregating account information across multiple banks/financial institution

  • Initiating payments across multiple financial institutions

  • Integrate with additional Fineract Open Banking APIs

  • Integrate with Mojaloop PISP APIs. 

  • Leverage completed components of UI library

Helpful Skills

Android development, SQL, Java, Javascript, Git, Spring, OpenJPA, Rest, Kotlin, Mojaloop

Impact

By providing an extensible open banking fintech app framework, allow partners a complete stack of Open Banking APIs and reference front-end application to rapidly build innovation via Open Banking APIs.   

Other Resources

Source code: https://github.com/openMF/pisp-app

Figma Design Mockups: https://www.figma.com/proto/ZkQJHzXWfjZugpTtz1MFdt/Mifos-PISP-App?node-id=10-27&starting-point-node-id=101%3A32&scaling=scale-down&t=vtBsVklbforA4MiB-1

Previous Development:

Google Whitepaper on 3PPI: https://static.googleusercontent.com/media/nextbillionusers.google/en//tools/3PPI-2021-whitepaper.pdf

UK Open Banking API Standard: https://standards.openbanking.org.uk/

Open Banking Developer Zone: https://openbanking.atlassian.net/wiki/spaces/DZ/overview

Examples of Open Banking Apps: https://www.ft.com/content/a5f0af78-133e-11e9-a581-4ff78404524e

See https://openmf.github.io/mobileapps.github.io/

Micro Front-ends Proof of Concept for Fineract 1.x & Mifos X

Mentors

 @Jose Hernandez @Aleksandar Vidakovic @Ed Cable

Category

Web - Mifos X Web App/Fineract

Overview & Objectives

Originally this idea was sparked by looking at what Moov has around UI drops -