I’m Justin. I’m a product manager at PrizePicks.
Helping members securely manage their account, stopping bad activity without getting in the way of trusted members, and fixing problems when the normal experience breaks down.
After three NFL seasons, I joined PrizePicks as an Associate Product Manager and built Platform Operations from the ground up. In 2026, I transformed Platform Operations into the Accounts domain.
2025 → 2026
I built Platform Operations from the ground up. The team grew substantially, with product work spanning support, settlements, and operational enablement.
In 2026, I moved into the core product organization and transformed Platform Operations into the Accounts domain. I separated the roadmap into three jobs with different member needs and tradeoffs.
HQ + legacy admin
One member issue could move through support, settlements, fraud operations, or incident response. Each team had part of the story, spread across multiple tools. At every handoff, the next person could spend time rebuilding what had already happened.
We built HQ to bring the history and the actions into one place. I mapped the workflows across those tools, designed the role-based access foundation, and led the use-case-by-use-case plan to move teams off the legacy admin safely.
Back-office working session
member asks about payout
entry still pending
review cleared
provider delay open
context copied here
action lives elsewhere
To answer one question, the operator had to rebuild the story.
About ¾ faster median time to close
About 40% faster median response time
RBAC role-based access foundation
Settlements
I tested whether game-end-to-payout time was associated with what members did next across major sports. The strongest relationship appeared in NFL. Settlement speed was worth treating as a product problem, not just a back-office timing issue.
The model showed a meaningful opportunity if we could reduce the wait. I took ownership in August 2025 and built the MVP settlement buffer: eligible entries could trigger once defined thresholds were met, without waiting for live scoring to mark the game final. The next milestone was confidence-based timing; we did not reach it before ownership moved.
The broader roadmap also delivered Win Protection, stronger resettlement controls, and a new Scoring Dashboard workflow.
Real-world game endProvider finalization sat inside the earlier wait.
Shipped MVP mechanism
The real-world event is over. The live-scoring provider may not have marked it final yet.
Move from fixed eligibility thresholds toward confidence in the underlying result.
About ⅔faster from game end to payout
2×+increase in entries settled early
Double-digitpoint gain in auto-settlement
Support + AI
FIN could explain a withdrawal policy. It could not answer “Where is my withdrawal?” because it did not have the member’s account context or a way to act.
I led the strategy with Customer Support, including the build-versus-buy decision, technical sprints, and a roadmap ordered around the member journey rather than integration ease.
40%+automation rate
About ⅔resolution rate
Double-digitCX score gain, YTD
Results shown as rounded ranges.Multi-accounting
Multi-accounting means one person operating more accounts than permitted. A large backlog had already been flagged, while the behaviors creating new accounts were still active. I authored and pitched a phased plan that treated those as two different jobs.
guardrailUnder 1% false positives
Across the back-book work, we held false positives to well under 1%. At large scale, even a small error rate could block legitimate members.
I spent three NFL seasons with the San Francisco 49ers, New York Giants, and Kansas City Chiefs. I’ve also advised early-stage teams through Ohio Angel Collective and Rootnote.
I also earned an M.S. in Consumer Sciences. My graduate work included consumer-insights research into customer pain points and habits, along with research on switching costs, satisfaction, and loyalty.
While football was still the job
Sleep stages, HRV, training load, nutrition, recovery habits, and my own read on how I performed—45 fields in all. I was trying to learn what actually helped, not build a portfolio project.
Each mark is a day I logged at least one signal. Missing days stay missing; nothing is filled in.
One person’s self-tracking data. Ratings such as RPE and overall performance are subjective. Missing values are shown, not filled. Relationships are descriptive—not causal or training advice. The source does not document the overnight sleep-date convention.