Business analysis and requirements work for a mortgage loan audit platform — competing-design tradeoff analysis for blended Credit/Compliance checklists, adaptive checklist filtering, a loan-info pop-out panel, and an RPA-to-API migration analysis across BotGenius clients.
A solo first-year master's project. HobbyHub is a low-intensity discovery platform that helps people explore new hobbies without the firehose of TikTok or YouTube. Built around HCI design-thinking — empathize, define, ideate, prototype, test — with a calm-mode interface, curated starter paths, and a guided onboarding quiz designed to reduce decision fatigue.
Led an integrated study application from concept through full product plan: WBS, business case, semester timeline, stakeholder & risk analyses, and cost models — building a strategic roadmap for collaborative academic support across a campus.
User research on Tinder pain points, synthesized into product opportunities. Designed two low-fi and one high-fi Figma prototype featuring a matchmaking quiz — projected to lift conversation quality 45% by prioritizing meaningful connections over match volume.
Reimagining Tinder around quality matches over quantity — a compatibility quiz and AI-assisted conversation starters built to help people connect on something real.
The purpose was to make focus-quality matches, not quantity matches, on the app Tinder. After collecting data on where the existing experience fell short — endless swiping, shallow profiles, and conversations that fizzled before they started — our team brainstormed how we could implement a more intentional matching feature directly inside the app.
The final prototype included a compatibility quiz that recommends people based on similar answers across career, lifestyle, and personality dimensions. The feature also implemented Artificial Intelligence to help new matches find interesting and unique topics to spark a conversation — so the moment after a match isn't a blank screen, it's an opening.
AuditGenius is a mortgage loan audit platform used by Credit and Compliance auditors to review loan files against configurable checklists. I drove requirements and backlog development for a UI Makeover initiative — translating operations-team brainstorms into groomed user stories, tradeoff analyses, and Jira-ready tickets.
The work spans end-to-end business analysis: eliciting requirements, structuring competing design options for decision-making, and managing open questions through backlog grooming. In parallel, I analyzed a large set of automation candidates on BotGenius, classifying viable bots for immediate API migration and pinpointing several high-leverage integrations to accelerate the RPA-to-API transition.
When a loan needed both a Credit and Compliance audit, stakeholders were split between two competing designs. I wrote both as parallel user stories with full acceptance criteria and a side-by-side tradeoff matrix, then isolated the "must resolve first" decisions (loan-release rules, lock-contention model) blocking each path.
Auditors were answering questions that didn't apply to the loan in front of them. I scoped the story to extend existing filtering logic rather than rebuild it, wrote acceptance criteria for edge cases like missing data and hidden questions, and produced a Jira-ready story with dependencies and open questions assigned to Ops.
The loan "More Info" panel competed for screen space with the checklist. I defined requirements for a pop-out panel — launch control, true window dragging, live data sync — and anticipated multi-window edge cases plus a compliance constraint that panel content must not be printable or savable.
Ramped up on the platform and Jira through shadowing, training, and feedback sessions with the team — including epics/stories/tasks/bugs structure, sprint-sized story breakdown, and writing testable acceptance criteria — while reviewing a competitor platform for context.
Drafted the first Blended Model checklist user stories and revised them from manager feedback, built an AuditGenius UI mockup (documented as both a doc and a slide deck), started the BotGenius bot-by-bot migration review, and picked up a new allocation-queue enhancement work item.
Split the reverification/checklist redesign into two focused stories — a configurable reverification pick list and expanded adaptive filtering — and pushed the BotGenius migration analysis (process, API feasibility, effort, and architecture per bot) through the large majority of the bot portfolio.
Reworked the shared loan-data-fields story for simultaneous Credit and Compliance audits, resolving the core architecture question in favor of one data model with two checklist views, then drafted the grooming document for adaptive checklist filtering.
Finalized the "More Info" pop-out panel story and a second, competing-design checklist story, and moved both into Jira for cross-team review — bringing AuditGenius, BotGenius, and the Blended Model redesign toward clearer implementation plans.