Selected Work

Selected Work & Approach.

Examples from institutional operations and workflow improvement, alongside independent and academic projects.

Professional Work & Approach

Selected Work & Approach.

Examples from institutional operations
and workflow improvement.

Workflow improvement / BNY

AI-Assisted Payment
Correspondence.

Built two enterprise AI assistants to support payment correspondence and requests for information.

The Need: Preparing consistent correspondence required repeated extraction, drafting, and checking.
My Contribution: Developed assistants on BNY’s enterprise AI platform to structure information and prepare drafts. Tested RFI scenarios with synthetic inputs and retained mandatory human review.
The Result: Both assistants were reviewed and approved at VP level for production use.

Process understanding → practical automation

THE WORKFLOW

Structure the information

Identify the fields a useful draft needs.

Prepare a consistent draft

Use the assistant for repeat preparation.

Keep human judgment

Review the content before it is used.

A high-level process summary.
No internal prompts, documents, or client data.

Personal work summaries. No client names or confidential transaction details. Views are my own.

Interactive Example

Reconciliation,
in Practice.

Reconciliation means comparing records, understanding what differs, and deciding what to investigate. Try a simplified example.

Illustrative demo · Fictional records · Runs in your browser

Choose a scenario

Comparison of two fictional financial records
Field	Operations	Counterparty
Reference	FX-1042	FX-1042
Currency	USD	USD
Amount	125,000.00	123,750.00

Difference: $1,250.00

The references match, but the amounts differ. Check the underlying activity and any fees or adjustments before deciding what to correct.

NEXT STEP / INVESTIGATE THE AMOUNT

This example explains a method. It does not connect to a financial system or reproduce a BNY tool.

Independent & Academic Projects

Analysis into Application.

View the Project Summaries

06 / Independent & academic projects

Exploring Finance & Technology.

Explore all projects

Independent case study

Global Reimbursements

Launch-readiness analysis covering partner oversight, payment reversals, pilot criteria, and auditable decisions.

Explore the case study

Cisco Smart Campus advisory-report cover

Academic group project · Erasmus University

Smart Campus Strategy

A proposed Cisco-enabled approach to campus security, hybrid learning, space utilization, and sustainability.

Read the project

Academic group project · Florida State University

Billboard Trends & Inflation

Exploratory data preparation, visualization, clustering, and predictor-importance analysis using Tableau and Tableau Prep.

Explore the analysis

Academic group project · Florida State University

AI Image Classification

A Teachable Machine model predicting hobby categories from webcam input, deployed in an original student web project.

Explore the prototype

Tools & methods

SQL · Tableau · Tableau Prep · Workflow analysis · Reconciliation · Enterprise AI assistants

Professional development

BNY AI for Everyone · L1 Explorer & L2 Builder

01 / Independent case study

Global Reimbursements

Assessing what a cross-border payment product needs before it scales.

The Question: How can a proposed global reimbursements feature expand internationally while maintaining oversight of partner-led payment processing?
The Approach: Reviewed partner accountability, dispute and reversal exposure, waitlist access criteria, and the auditability of payment decisions.
The Recommendation: A gated launch with defined oversight, settled loss allocation, structured exception records, and evidence-based criteria for expansion.

Global paymentsProduct analysisOperational risk

Read the case study

Independent analysis of a proposed feature. Recommendations are case-study judgments.

02 / Erasmus University · Academic group project · 2024

Smart Campus Strategy

Connecting university needs to a proposed technology implementation plan.

The Question: How could Erasmus University use connected-campus technologies to address security, learning, space utilization, and sustainability?
The Approach: Assessed campus stakeholders and challenges, then evaluated Cisco IoT, network-security, access-management, and smart-workspace solutions.
The Recommendation: A phased pilot and evaluation plan, followed by wider implementation, training, feedback, and ongoing monitoring.

Technology strategyStakeholder analysisImplementation planning

Read the advisory report

Academic group proposal developed during my Erasmus residency. The report does not document a deployed Cisco or university implementation.

Page from the Billboard and inflation project showing a predictor-importance chart targeting Spotify popularity

03 / Florida State University · Academic group project · 2024

Billboard Trends & Inflation

Exploring relationships between economic conditions and the characteristics of No. 1 songs.

The Question: What patterns appear when Billboard chart history, music attributes, and annual inflation are compared across decades?
The Methods: Used Tableau Prep to clean and combine datasets, Tableau to visualize trends, K-means clustering to compare song profiles, and predictor-importance models targeting Spotify popularity and annual inflation.
The Data Work: Removed repeat chart-week entries, aligned records by year, and examined tempo, energy, danceability, valence, and popularity.

TableauTableau PrepK-means clustering

Read the project report

Exploratory academic analysis. Associations do not establish causality; the original project used a partial-year estimate for 2024 inflation.

04 / Florida State University · AI seminar

Hobby Image Classifier

A browser-based introduction to training and deploying an image-classification model.

The Project: Built and deployed a Teachable Machine image classifier with a team during an AI seminar at FSU.
The Interaction: The model predicts hobby categories from live webcam input. The original project is hosted on Weebly.
The Focus: Turn a trained model into an interface people can try, while observing how the input affects the model’s predictions.

Teachable MachineImage classificationPrototype deployment

Open the original project

Opens the original student project on Weebly. Webcam access, if requested there, remains controlled by your browser.

Interested in how this work connects to institutional finance?

View my professional experience

For a conversation or further information.

Contact Me

Selected Work & Approach.