I work upstream of the build. I frame the real user problem, fit AI into the workflow, and shape how the product is positioned. The two projects below show how I think, not just what I shipped.
I'm a marketing and product strategist with an MS in Marketing from Johns Hopkins Carey and a background in planning at McCann Worldgroup. What sets me apart is how I use AI: to think better, not just produce more. I use it to turn ambiguity into clear workflows, pressure-test ideas, and keep human judgment on the final decision.
Most products answer the wrong question well. I start with the user's real friction and the assumptions everyone skipped, then work out where AI fits, how to route the work, and how to position the result so its value is obvious to the person choosing it.
SubletU and FOVRA are the two halves of how I work. One is a two-sided marketplace built on a concrete, under-served pain point. The other is a system for making AI dependable enough for serious work. Both started as a problem I could name before I could solve it.
Structuring research, drafting, critique, and QA into routed workflows, not one-shot prompts.
Scoping the MVP to its wedge and resisting complexity that doesn't earn its place.
Naming the real friction and the unspoken assumptions before committing to a build.
Making the value impossible to miss for the person who has to choose it.
A student subletting platform that closes the gap between 12-month leases and a 9-month academic year.
Leases run 12 months; school runs about 9. Students either eat the cost of summers they're not there, or scramble to sublet through chaotic group chats where listings are inconsistent and scams are common.
Sublets already exist; what's missing is a verified, legible channel. Students move quickly when they trust the counterparty. The right move is to standardize and verify, not add another open marketplace to scroll through.
SubletU keeps the network to verified students, standardizes how every sublet is listed, and cuts the time it takes to match. A stressful, scam-prone scramble becomes a clean, predictable process.
Framed the core problem, defined the MVP, made trust-and-verification the wedge, and used AI to pressure-test the user problem and sharpen positioning for a student audience.
The hard part of a marketplace isn't listings. It's getting strangers to trust each other quickly. I built the product around that single problem and relied on verification, structured listings, and reviews instead of a long feature list.
The verified-student angle wasn't a guess. It came out of a structured, AI-assisted research process that pressure-tested the idea, with human judgment on the final call, before any product was built.
Instead of asking AI for positioning directly, we first mapped the functional, emotional, and social needs that Craigslist and Facebook Marketplace leave exposed for students. From there we generated six positioning concepts and chose the trust layer ourselves.
We built digital twins of two student segments and ran qualitative research on feature trade-offs, messaging resonance across four angles, and competitive fit through a five-dimension brand similarity matrix.
She has a signed lease and is leaving for study abroad or an internship. Her real barrier is not knowing who's contacting her, so she won't post until that's solved. She responds to trust and safety.
Arriving for a summer internship under deadline pressure. His pain is incomplete listings and hours of back-and-forth to extract basic facts. Responds to speed & information quality.
Uncontested white space. Across five dimensions drawn from the segments, SubletU was the only concept to score above 3 on both trust and student relevance at once. That's a position no existing platform holds. Phase 1 targets the Baltimore and DC corridor (Johns Hopkins, UMD, Georgetown, GWU), where demand is concentrated and predictable.
An AI workflow and QA system that makes AI dependable enough for serious, high-stakes work.
People use one model, one prompt, one pass, then trust the output. For anything that matters, that's fragile. Weak assumptions go unchallenged, the wrong model gets the wrong job, and nothing catches what's confidently wrong.
Serious thinking separates research, writing, critique, and review into stages, each with the right tool and a check on the last. AI should work the same way. What sets it apart isn't generation, it's the QA layer.
FOVRA breaks AI work into routed stages (research, writing, critique, refinement, and final QA) instead of one long chat. It helps users pick the right model for each stage, challenge weak assumptions, and review every output before it's trusted.
Set the thesis (structure beats single-shot prompting), designed the stage-based workflow and QA model, and positioned FOVRA away from generic "AI automation" language.
Everyone is racing to generate faster. I focused on the opposite end of the workflow: the step that decides whether an output can be trusted. Making QA the core feature is what turns FOVRA into a defensible category rather than another AI chat tool.
One is a marketplace, one is AI infrastructure, but they come from the same playbook: find the real problem, design a structured solution, and obsess over what makes it trustworthy.
SubletU started from a calendar mismatch; FOVRA from why single-shot AI fails. I lead with the real friction and the assumptions everyone else skips.
Both win on structure: verification and standardization in one, stage-routing and QA in the other. I design the workflow, not just the surface.
Reviews and verification in SubletU; a dedicated QA layer in FOVRA. I make what earns trust a first-class feature, because that's what makes a product credible.
I prioritize the one thing that has to be true for the product to work (trust for SubletU, structured QA for FOVRA) and cut the rest.
I can explain why a product is different in language the buyer actually uses, which keeps FOVRA out of the generic "AI automation" category and SubletU out of "another listings app."
I use AI to move fast and think more clearly, routing each task to the right method while keeping a person in control of the final decision.
I'm drawn to roles where framing the problem matters as much as executing the answer. If that's how your team works, I'd love to connect.