I work across business analysis, operations, and AI-assisted projects. I like building things, picking up new skills, and exploring where AI can be useful.
This is a collection of what I’ve worked on, what I’ve tried, and what I’m learning along the way.
OPEN TO EARLY-CAREER BA & OPERATIONS OPPORTUNITIESAI AGENT DESIGN · REQUIREMENTS · PROTOTYPING
01 / SELECTED WORK
Projects that show how I investigate, coordinate, decide, and test.
01OPERATIONS ANALYSIS2025
Amazon Improvement Proposal
Independently analyzed alarms and operational data during a two-week project as an Area Manager intern, then presented a conveyor-relocation recommendation and pilot plan.
A mobile workout app shaped by my own training. I use AI-assisted development to turn workout notes into reviewable routines, simplify set logging, and track progress through working weight, reps, and volume.
EARLIER APP SCREENS · AI routine intake & workout logging Select a screen to view full size.
03SMALL-BUSINESS OPERATIONS2020–22
DvzzPins
Turned product ideas into manufacturing specifications, reviewed supplier proofs, coordinated restocks, and questioned repeat costs while operating an e-commerce storefront.
SUPPLIER PROOF · Specifications reviewed with GS-JJ · View full size ↗
04PROJECT COORDINATIONTEAM PROJECT
CCNY Capstone
Served as the primary project manager for a student team developing recommendations for communication, belonging, and student engagement.
CONCEPTUAL ROLE MAP · Based on my responsibilities and the team presentation.View full size ↗
02 / HOW I WORK
I learn by doing. The work starts with the problem in front of me.
01
Understand the friction.
Look for what is unclear, manual, mismatched, or creating avoidable effort.
02
Make the work visible.
Turn observations into a plan, a requirement, a decision, or a prototype people can react to.
03
Test and refine.
Test AI-generated work, challenge unsupported outputs, and refine instructions and prototypes through hands-on use.
03 / AI AGENT EXPERIMENT
An AI agent that asks the next question.
I designed an AI agent to help me think through business analysis. I shaped its instructions to challenge vague requirements, ask about exceptions, and help turn my answers into user stories and testable acceptance criteria. I review the outputs and keep unresolved decisions visible.
from-idea-to-requirement
ILLUSTRATIVE SCENARIO Designing an expense-claim feature. Fictional dialogue showing the discovery approach I designed for the agent.
DAVID
The app should automatically approve expenses under $50.
AI AGENT
Is the amount enough to decide? What if the receipt is missing, the expense is personal, or it’s a duplicate?
DAVID
Require a receipt and a business purpose. Send possible duplicates to a reviewer.
AI AGENT
Should the system approve the claim, or check whether it’s ready for a manager? Those need different permissions and acceptance criteria.
DAVID
Keep manager approval. My goal is less manual checking, not a new approval policy.
AI AGENT
Then define a pre-check, with exceptions routed for review. Here’s a draft to confirm with the policy owner before implementation.
See the proposed user story & acceptance criteria +
User story
As an expense approver, I want claims pre-checked for completeness and possible duplicates so I can focus my review on exceptions and business justification.
Acceptance criteria
Ready for review: Given a claim under $50 with a receipt, a business purpose, and no duplicate flag, when the pre-check runs, then mark it “Ready for review” and keep approval pending.
Missing information: Given a claim under $50 with no receipt or business purpose, when the pre-check runs, then mark it “Needs information,” identify what is missing, and keep approval pending.
Possible duplicate: Given a complete claim under $50 that is flagged as a possible duplicate, when the pre-check runs, then route it for manual review and keep approval pending.
Still to confirm
Who owns the policy? What counts as a duplicate? What is the review path for claims of $50 or more? A recorded business purpose alone does not prove policy eligibility.
04 / ABOUT
A bit more about me.
At Computershare’s Canton, MA office, representing FDM and supporting the Copilot Hackathon.
I enjoy learning new hobbies and building things. Right now, I’m especially interested in AI and finding more ways to use it. I like showing friends what I’ve learned and helping them try AI for themselves.
Music is a big part of my life. I used to produce music and sold beats online in 2020. Fitness is another interest, and it led me to build TetsuGym around my own training.
At Amazon, I provided first-line IT support, troubleshooting warehouse hardware, documenting equipment issues, and escalating incidents. At the Computershare Copilot Hackathon, I supported participants with Copilot Studio workflows, connectors, and environments, plus Power Automate integrations.
