I rebuild the data systems
people depend on,
then put AI on top of clean data.
Product manager at Newmark. Former software engineer.
I’m rebuilding Newmark’s 20-year-old CRE research platform. Phase 1 has been live in NYC since October 2026, with the other six markets rolling out in phases. The platform serves clean research data to Perplexity over MCP. I own the data model, requirements, QA and roadmap for a three-person engineering team.
WorkFrom the business case to the data model
I work close to the build. At Newmark, fixing the legacy platform and interviewing about 20 analysts and managers showed where the data structure was failing. That work became the replacement plan and roadmap. For phase 1, I chose not to build a new retrieval or mastering model, so the cutover could not break the research analysts already run. I own the ERD, database design, QA and Scrum, and use Cursor for prototype screens, the data remodel and migration tests.
Phase 1 of the rebuilt CRE research platform has been live in NYC since October 2026; the other six markets release in phases. The platform serves 80+ analysts and 10+ managers across seven US and Canadian markets.
Before approval, from August 2025 to May 2026, I built the first bulk import and bulk update, fixed about 15 reports, tightened data access on about 67 reports, and interviewed about 20 analysts and managers. That work shaped the replacement plan and roadmap. I presented the business case to the CIO; with CIO sponsorship, the CAO approved a $650K phase 1 in May 2026.
I designed the data model across space, building, submarket and market, with lease and comp categorization and dirty data blocked at ingestion. Clean research data is served over MCP. About seven NYC analysts and managers use Perplexity on it daily for comps analysis and to feed research back into the platform as data updates. I own the roadmap, requirements, ERD/database design, QA and Scrum for a three-person engineering team.
Worked with 15+ small-business clients referred through municipal economic-development programs on market research, segmentation and pricing. Repositioned Bhlen, a digital guides business, through segmentation and messaging workshops aimed at decision-stage buyers. Ran outreach and five customer interviews for Revby’s planned digital service; the findings set its service tiers and pricing.
A supplier reporting workflow kept producing bad file output. I read the existing code with supplier teams before scoping, then rebuilt the workflow instead of patching it again. Outages fell from about 10 a month to 2.
Reworked purchasing approval APIs and support tools with Finance and Procurement, cutting procurement turnaround time by 30%. Ran backlog grooming, sprint planning and release coordination for platforms used by global suppliers.
Built Salesforce procurement automation and a trackable purchase order system.
Projects
From interviews to a working rental operations prototype.
Four landlord and tenant interviews, a PRD and user stories, followed by n8n/MCP automations for rent reminders, receipts, rent-increase notices, lease-expiry notices and missed-payment alerts. Prototype; no users yet.
Four principles. Both lenses.
Strong product strategy is rarely about adding more. It's about removing noise, sharpening the problem, and making the right tradeoffs visible. I like product work that gets clearer as it gets simpler.
A product no one understands will struggle to move. I treat communication as part of the product itself — discovery findings, a roadmap, a PRD, a recommendation. The narrative has to help people believe in what comes next.
Useful AI depends on the data it can access. I start with the entities, relationships and ingestion rules, then connect that data to the workflows people need.
I start by watching how people work and asking where the system gets in their way. At Newmark, legacy fixes and interviews exposed the data problems that shaped the replacement plan.
I find the real problem before the team commits to the wrong solution — interviews, shadowing, workflow audits, synthesis.
I design the data model behind MCP access and build workflow automations with n8n. Clean data gives AI a useful foundation.
I use data to clarify tradeoffs and make product conversations concrete — the kind that help teams act, not dashboards that just look polished.
I turn ambiguity into clean PRDs, structured backlogs, and plans teams can trust. Clarity reduces waste and rework.
I bring strategic framing to product choices — prioritization, vendor evaluation, build-vs-buy, aligning direction with business reality.
I connect technical detail with business understanding — helping different stakeholders see the same picture clearly enough to move together.
LET’S
BUILD
SOMETHING WORTH
CARING ABOUT.
I’m looking for applied-AI and forward-deployed product roles where I can work with users, shape the data model and build alongside engineering. I’m also open to technical PM roles on AI and data platforms.
Open to — Applied-AI / Forward-Deployed PM · Technical PM, AI/Data Platform
Based in — New York City · open to NYC, remote, and select relocation
Work authorization — Authorized to work in the U.S. on STEM OPT through 2028; will need H-1B sponsorship