IS484 IS Project Experience (FinTech)

Course Description:

• This is an SMU-X course designed in collaboration with participating Banks, FinTechs, and other FIs, to serve as project sponsors. Collectively, industry sponsors will supply a minimum of 5 projects ideas to select from.
• Students will form teams of 5 or 6, and select one the project ideas to work on. Project selections do not need to be unique, meaning multiple teams can select the same project idea.
• Each student project team will be assigned to a sponsor/mentor and an SMU faculty supervisor.
• Sponsors will provide project scope and management for student teams to have practical industry learning experiences.
• Student teams will have weekly check in meetings, either virtually or physically, with their sponsor.
• Sponsors will specify the technologies to be used, including; development tools/languages, OS, database, 3rd party libraries, target deployment environment e.g. cloud environment.
• Student project teams will be expected to develop a working software application prototype, to be delivered to the sponsor at the end of the course.

Course Prerequisites:

1. Software Project Management (IS212) is a pre-requisite or a co-requisite.
2. Any two (2) track courses from the track that you are declaring for your project. One of these courses can be a co-requisite.

Project Timeline:

Activities	Timeline	Term 1/ Term 2	Action By
Project Sourcing and Registration	Week -14 to Week -10	Form teams. Review the below set of predefined projects provided by Citibank, OCBC, NETS, UBS, and others. Fill up the Project Team Signup Sheet at the below link, listing your preferred projects. FT Track Coordinator will finalize the matching of teams to projects.	Students
Project Matching	Week -10	FT Track Coordinator will finalize the matching of teams to projects.	FT Track Coordinator
Proposal	Due before the start of Week -8	Submit your project proposals to your Track Coordinator(s). For mixed-track teams, both track coordinators need to review your proposal.	Students
Decision on Proposal	Week -4	Your Track Coordinator(s) will confirm that the project has sufficient scope to fulfill your respective track requirements for IS Project Experience.	Track Coordinator, Students, (Optional: Sponsor)
Start of Project	Week 1	Supervisor - Teams	Student
Midterm	Week 8	Presentation	Students, Supervisor, Reviewer (Optional: Sponsor, Track Coordinator)
Finals	Week 14 to Week 16	Presentation	Students, Supervisor, Reviewer (Optional: Sponsor, Track Coordinator)

Key Dates for AY2025/26 Term 1

• ✓Students submit team registration and prerequisites forms: by 13 Apr 2025
• Sponsor project descriptions available: by 2 May 2025
• Sponsor info session with students: week of 5 May 2025
• ✓Students submit project preferences: by 14 May 2025
• Student teams assigned to projects: week of 19 May 2025
• ✓Students submit project proposals: by 12 Jun 2025
• SMU track coordinator approves proposals: around mid-Jul 2025
• Students start working on projects: in early Aug 2025
• Midterm review: around mid-Oct 2025
• Final presentation/demo: around the end of Nov 2025

Project Team Signup Sheet:

AY2025/26 Term 1
https://docs.google.com/spreadsheets/d/1q-2qNkXGcjPxybU52s-1cazP5k4zhHTYRn7SKxz5Hjg/edit?gid=0#gid=0

Current Projects - FY2025/26 Term 1

Proj #	Sponsor Organisation	Title	Project Mentor	Business Domain	Problem Statement	Project Description	Project Inputs	Project Deliverables	Fulfills BA Requirements
1	Tiger Fund Management	News/Forum Sentiment Analysis for Investing	Gabriel Woon	Investment Management	Tiger Fund Management is a Singapore-based investment management firm specializing in active asset allocation, using a mix of quantitative strategies with fundamental equity research. Our core mission is to deliver superior risk-adjusted returns through a disciplined and systematic investment process. The firm manages a diversified portfolio spanning public equities and fixed income, with a strong emphasis on technology-driven decision-making and data-informed strategies. In today’s fast-moving financial markets, investor sentiment plays a significant role in influencing stock price movements. However, traditional investment analysis often overlooks qualitative signals hidden in unstructured data sources—such as earnings call transcripts, news articles, and analyst commentaries. Tiger Fund Management is exploring how sentiment analysis can be applied to enhance investment decision-making. The challenge is to develop a sentiment analysis pipeline that can extract and interpret sentiment signals from financial text data and assess their impact on stock performance. This tool should help investors better understand market mood, anticipate potential movements, and support portfolio strategy decisions.	1. Goal The goal is to build a working and demonstrable stand-alone proof-of-concept system that performs sentiment analysis on financial text data to support stock investment decisions. This project will explore how natural language processing (NLP) can be applied to extract and visualize sentiment signals from related market commentary. 2. Background and Scope Investor sentiment is an increasingly important factor in financial markets, especially in interpreting qualitative cues from news articles, and analyst reports. This project, sponsored by Tiger Fund Management, aims to create a sentiment analysis tool tailored for the investment domain. Students will research financial sentiment models, extract sentiment from unstructured text, and correlate it with stock price movement. The scope includes: - Designing a sentiment extraction pipeline using NLP techniques (e.g. FinBERT) - Correlating sentiment trends with stock returns - Building an interactive dashboard for investment insights 3. Data Ingestion Requirements - News headlines or analyst commentary for additional sentiment context (e.g CNN, Reddit, Bloomberg) - Historical stock price data (via Yahoo Finance) 4. Data Analysis Requirements - Sentence-level sentiment classification (positive, neutral, negative) - Time-aligned comparison between sentiment score trends and stock price movement - Correlation or regression analysis to assess predictive power of sentiment 5. Dashboard/UI Requirements - Summary view showing sentiment trends across stock price and time periods - Exportable results (e.g. Excel or CSV)	The sponsor will share a simple earnings call sentiment analysis.	1. Sentiment Analysis Pipeline A working Python-based pipeline that performs sentiment classification on financial text (e.g news and forums) using pre-trained models such as FinBERT. 2. Data Processing Framework Scripts or workflows for ingesting and cleaning text data, parsing and splitting transcripts by section, fetching historical stock price data for comparison. 3. Sentiment Scoring Output Exportable sentiment data (e.g., CSV or Excel) showing sentence-level sentiment, aggregated sentiment by section, timeline alignment with stock price movement. 4. Interactive Dashboard A user-friendly dashboard (e.g., using Streamlit, Power BI, or Plotly Dash) to visualize sentiment trends over time, compare sentiment across companies or speakers, and link sentiment to stock returns.	Yes
2	Viom Technology Services	Business Process Mining and Improvement	Amit Gupta	Finance	Financial institutions rely on complex workflows for activities like customer onboarding, loan approvals, and fund transfers. Process mining helps reveal real process flows vs. intended ones, process management optimizes them for efficiency, and automation reduces manual effort and compliance risks. In an evolving landscape like Open Banking, having real-time visibility, process traceability, and continuous improvement is critical for operational agility and regulatory compliance. This prototype focuses on helping financial institutions mine, manage, optimize, and automate their core processes to boost efficiency and meet compliance standards.	The goal of this assignment is to design and develop a lightweight system that first performs process mining based on configurable sources (such as logs, events, or databases) and enables process management once the process is mined. The system should also include a Process Repository to store, track, and retrieve processes for management purposes. This prototype should leverage Generative AI or an LLM to provide recommendations for improving processes, enhance decision-making, and automate workflows. Additionally, the system should enable process visualization, bottleneck detection, KPI monitoring, process optimization, and compliance monitoring. The prototype should be capable of generating syntactically valid BPMN models and visually displaying a process map with improvements.	Project sponsors will share sufficient context so students can understand the scope, design, tooling, and expectations of the prototype. The mock raw data files, explanation of this data structure and other useful details can be shared.	Prototype: A working prototype (including source code, configuration files, process maps, and BPMN exports). Documentation: A detailed report that summarises research findings and explains the design, development process, challenges, and solutions for the prototype.	Yes
3	UOB Kay Hian	Research and Investment Automation	Sherry Yang	Research and Investment	The department handles the research and investment of a variety of asset classes especially equity thus we need to do the results’ review for many companies and reflect our position changes from time to time. We are looking for automation help to improve the working efficiency and accuracy. Manual handling of data collection and stock transaction records is inefficient and prone to human error. The lack of automation limits responsiveness and increases operational burden for small teams.	Task A: Automated Financial Data Extraction & Template Filling The current process of retrieving financial information (e.g., company earnings, balance sheets, macroeconomic indicators) involves manually browsing websites such as statistics departments, and corporate investor relations pages. The data is then copied into structured Excel templates used for reporting and analysis. Task B: Automated Investment Portfolio Tracking Currently the investment team uses spreadsheets to manually update the market value of equity positions held by the principal. For each transaction, prices are manually looked up and gain/loss is calculated across various time horizons.	Data collection sources (financial reports and govt statistic info, etc); Manual Excel templates currently in use (to understand format & structure); Historical mock transaction data (buy/sell/stock code/quantity/price/date)	A working prototype (script, app, or dashboard) that automates: data scraping from selected websites or APIs; population of Excel-based templates; calculation of real-time or historical equity portfolio values Documentation includes user manual; codebase and deployment instructions; maintenance recommendations for future use.	TBD
4	Singapura Finance Limited	Board Dashboard	Loh Ching Soo	Lending / Corporate Management	Singapura Finance Limited (SGX:S23) is a Singapore-based finance company. The Company's services include personal savings, corporate deposits, and consumer and corporate loans. The Board of Directors meets regularly and via the various Committees (Audit, Risk, Digitalisation, Remuneration and Nomination). Reports are detailed, extensive and scattered via office productivity apps in Board.Vision and sometimes email. There is no easy way to get a consolidated snapshot of the backwards-looking and forward-looking indicators of the business.	A Board Dashboard presenting a clear, real-time view of critical company metrics across financial, funding, loan, risk, customer, strategic, and human capital areas. Intuitive and delightful UI/UX (some users not digitally comfortable), built on design thinking principles for both input/upload (manual data like Excel) and consumption. Draft Metrics: 1. Financial Performance 2. Funding and Deposit Health (CASA Focus) 3. Loan Portfolio Health 4. Risk and Compliance Metrics 5. Customer and Market Metrics 6. Strategic Initiative Progress 7. Human Capital and Governance	Data inputs – core banking, Excel, future CRM and future-proofed	Real-time dashboard with monthly and quarterly snapshots, health indicator (traffic light) and timeline trends. Compliance with MAS and SGX regulations, with a stringent focus on security, privacy and confidentiality. Private cloud, locally hosted. Modular backend for easy modification and deletion/addition of metrics, perhaps using no-code or low-code methods. Print and download functions. Mobile and web channels.	Yes
5	Citi	Deal Review Committee Using LLM Agents	Nirav Parikh	Investment Banking, Sales & Trading, Structured Products	Investment banks face challenges in evaluating deal proposals due to the need for multi-faceted analysis across revenue potential, risk assessment, regulatory compliance, and strategic fit. This project aims to develop a framework leveraging LLM agents to provide a holistic and expert-driven approach to deal review, improving both the efficiency and accuracy of decision-making processes.	In investment banking, evaluating deals for new clients involves multiple dimensions of analysis. Banks must assess the market potential, competitive positioning, and risk profile of a client's business, while estimating revenue and cost projections. Additionally, ensuring compliance with regulatory standards and aligning deals with the bank's strategic objectives are crucial aspects of decision-making. This project proposes the development of a framework utilizing Large Language Model (LLM) agents to simulate a virtual committee of financial experts. Each LLM agent will be specialized in specific areas such as risk assessment (credit, market, operational, regulatory, reputational), revenue estimation, compliance, and strategic alignment.	Students will be provided with: - Training data from previous deal reviews, including analysis from financial analysts and risk managers. - Access to relevant market data, risk factors, and financial models. - Strategic guidelines and risk appetite documentation for new deals. Note: Anonymized Citi internal data (no client data) will be required for this project subject to approvals and potential NDA agreements.	- LLM-Based Expert Agents: A framework of specialized LLM agents trained to simulate expert perspectives in areas such as revenue estimation, risk analysis, compliance, and strategic alignment. - Virtual Committee Decision Process: A mechanism for synthesizing the insights from different agents to form a comprehensive recommendation on deal approval and conditions. - Decision Support Dashboard: A user interface that provides deal recommendations, approval conditions, and risk mitigation strategies based on the committee's output. - Documentation: Comprehensive documentation outlining the design, methodology, and decision-making process of the virtual committee of LLM agents. - Presentation: A final presentation showcasing the framework, its decision-making process, and its potential impact on the bank's deal review process.	Yes
6	UBS	News Screener For Relevant Investment Opportunities	Hossain, Mohammad-Jahangir	Equities (Stocks)	Given current volatile world and lots of information being generated every moment, analysts require smart intelligent tools to sort through all those and provide clients with relevant investment advices on timely manner. The tool is to automate scanning of news, providing digest about the news in categories like sector, region and entity. Also to provide relevant sentiment of the news.	Bankers and Investment counselors (ICs) develop deep relationships with clients and provide them with relevant investment and opportunities advices based on client’s needs. The ask is to create a tool that: - Scans publicly available social media and news sources for news about relevant sector’s or region’s current and ongoing events and their co-related relevant companies or entities. - Summarize these views to a digestible format for Client Advisor - Ensure each data point has a source link Example: Tools scans news from the ‘Financial Times’ for relevant sector ‘Real Estate’ and region ‘China’, Based on the news coverage, it identifies the current market trend and affected companies and instruments relevant to those companies and provides a relevant view to client advisors.	- Project sponsors will provide 5 entity names, and suggested data sources on which the output should be created - Project Sponsors will review at regular intervals the outputs to refine requirements and usability of output - We will provide support on how to perform identity matching for the entities - Team can use AI tools to discover digest of news	- Working dashboard that provides a real view of relevant sectors and entities. - View provides contextual correlated sentiment assessment of entities. - View provides trigger notifications when significant activity threshold is breached (optional). - View provides collected historical information and perform system end to end risk and returns on entities. - Ability to collect data from different sources.	Yes

Archived Past Projects and Groups

https://docs.google.com/spreadsheets/d/1IDAhC4JiK3RuKnIDQMG5UjJ6I1IiImo81Lu13wAuUxE/edit#gid=491663198

Past Project Descriptions
https://docs.google.com/spreadsheets/d/1f7r2y1n6USWAYVTVyWkZkYxJt7HCoxlc-z3jAGzX7LQ/edit?gid=0#gid=0