COMMERCIAL JUDGMENT. HANDS-ON EXECUTION.
I bring product direction, strong teams, and hands-on AI execution to the same table.
CPO · VP Product · Head of Product New York City
Connect the customer problem to the business model. Decide what matters.
Set the direction, make the tradeoffs explicit, and give people room to deliver.
Build, test, and learn. Use AI where it helps turn a good decision into a useful product.
01 / THE DECISION ROOM
INTERACTIVE · ILLUSTRATIVETake a seat. Change the priorities.
See what a decision gives you,
and what it asks you to give up.
THE SITUATION
A hypothetical consumer product attracts new customers, but too few come back. You have a fixed effort budget. Where would you put it?
Move the slider to shift a fixed, hypothetical effort budget.
REPEAT-USE INDEX
127vs. a 100 starting point
THE TRADEOFF
More effort goes into repeat value. Near-term reach gets less investment, and the decision needs cohort evidence before you scale it.
Illustrative only. All values are invented to make tradeoffs visible. This is not a forecast or a record of company performance.
A model starts the conversation.
Customer evidence makes the decision.
02 / LEADERSHIP IN PRACTICE
HYPOTHETICAL CASES · NOT COMPANY HISTORYThree situations that reveal the work of product leadership: diagnosis, prioritization, and follow-through.
CHALLENGE
A hypothetical consumer product attracts new customers, but too few find a reason to return.
DECISION & TRADEOFF
Investigate the first useful moment. Shift some effort from acquisition to onboarding and repeat use, accepting slower near-term reach.
OUTCOME TO LOOK FOR
More customers reach value and come back. Validate by cohort before increasing acquisition again.
CHALLENGE
A hypothetical team is busy across several initiatives, while dependencies and shifting requests delay useful releases.
DECISION & TRADEOFF
Choose one customer outcome. Give it a clear owner. Pause attractive work that competes for the same capacity.
OUTCOME TO LOOK FOR
Shorter feedback loops and clearer accountability. Watch the risks created by the work you paused.
CHALLENGE
A hypothetical company has impressive demos, but no reliable path from a model's output to a customer's completed task.
DECISION & TRADEOFF
Narrow the use case. Define a quality bar, an evaluation set, and a fallback. Invest in reliability before widening the feature set.
OUTCOME TO LOOK FOR
A workflow customers can trust. Measure completed tasks, failure modes, and support burden before expanding.
03 / TRACK RECORD
I've led product at consumer platforms. Today, I also build and operate products myself. That keeps strategy close to the customer, the code, and the business.
Full background on LinkedIn (opens in a new tab)WAYS TO WORK TOGETHER
Full-time product leadership, fractional support, or a focused advisory engagement.
FULL-TIME
Own the product direction, build the team, and connect the roadmap to customer and commercial outcomes.
Discuss a leadership roleFRACTIONAL
Work alongside a team that needs experienced product leadership before making a full-time hire.
Discuss fractional supportADVISORY
Product audits, strategic sprints, growth systems, M&A diligence, and board-level product conversations.
Discuss an advisory engagementThe operating approach behind the work.
Connect acquisition, onboarding, repeat use, pricing, and measurement. Use channel automation across Reddit, Google, TikTok, Meta, and Instagram where it makes the work more useful.
Define the customer task, choose the model for the job, and keep architecture, evaluation, and review in the loop. I stay hands-on with the code and the product decisions.
Started at Dell EMC. Learned enterprise M&A integration before most people knew what product management was. That foundation still pays.
I'm a first-generation immigrant, a Queens College graduate, and a Baruch MBA. Building useful tools and giving back to CUNY remain part of the work.
Read the Extern case study04 / SELECTED BUILDS
STRATEGY MEETS THE WORKA few products from the portfolio.
Different problems. The same appetite
for making something useful.
Virtual staging and photo tools for property listings.
Restaurant inspection data, made easier to find and understand.
Monitoring and automated reservations for Too Good To Go bags.
Virtual staging and photo tools for property listings.
Monitors Too Good To Go for favorite surprise bags and supports automated reservations.
Restaurant inspection information, with a mobile augmented-reality view of nearby grades.
A shared canvas widget for your friend circle. Doodles and photos land on everyone's home screen in seconds.
Public idea boards, honest voting, and a private line to the founder. Customer messages land in Discord, while roadmap changes reach the people who asked for them.
Free tools that help people draft demand letters for disputes with landlords, contractors, and businesses.
Real guidance from people who have been there, with guide fees directed to nonprofits.
Upload a pet photo and explore a range of AI portrait styles.
Live watts and cadence from any Bluetooth power meter. Offline, clean, CSV export. My only Flutter app.
Proactive construction QA marketplace. Async pro reviews, dispute-ready PDF.
One-link family updates, no accounts. Built for the "how is grandma doing" use case.
Mets fan rage in 1996 GeoCities form. Live W-L from the MLB API, real payroll, a global fire-this-guy counter, a voting board for fan demands. Marquees, blinks, broken visitor counter. A joke, and also serious, and also sad.
WireGuard VPN done right. Invitation codes. QR setup. Self-hosted.
The Iranian classic, online. Play, track, keep the tradition alive.
Deploy and manage AI assistants from one dashboard.
One shared list for couples who split the Costco run. Real-time sync, no more duplicate ketchup.
Multi-protocol VPN distribution via Telegram for censored regions. VLESS+Reality, CDN-routed WebSocket, Shadowsocks with auto failover. Zero-knowledge.
Multiplayer flight combat over a photoreal 3D Manhattan. Real-time WebSocket dogfights.
Landlord ops hub with audit-defensible hour tracking for the real-estate-professional tax status.
A modern remake of the 1988 classic Life & Death. Diagnose, cut, and try not to kill the patient. Offline.
A love letter to SimTower. Stack the floors, run the elevators, keep tenants happy. On-device sim.
A calm companion for vet questions, everyday care, and outdoor conditions—all in one place.
Two-sided marketplace where brands recruit micro-influencers and AI ranks every applicant. Creators paid via Stripe Connect.
05 / NOTES FROM THE WORK
Writing on building, leadership, and the decisions behind the product.
Aug 21, 2026 · 6 min read
BagRescue earned media coverage with proprietary market data. The more important lesson: knowing which impressive claim to delete is part of the product.
Read →Mar 23, 2026 · 3 min read
I was quoted in the New York Post on AI and privacy. The hard truth: data is the asset, convenience is the pitch, and nine-figure fines are rounding errors.
Read →Mar 16, 2026 · 3 min read
I've always been the person who signs up for a product and immediately starts thinking about what I'd change. So I started writing it down.
Read →THE NEXT CHAPTER
Let's talk about product leadership.
Advisory conversations welcome, too.