Ben Siverly
Senior Product Manager · AI-Native Products
benjaminsiverly@gmail.com
linkedin.com/in/be-ns
github.com/be-ns · bensiverly.com
Summary
Data scientist turned product manager. I own the rental application product line at AppFolio, 8–9M applications a year across 20k+ property managers, with sole responsibility for strategy and execution. Outside work I build Brava, a recruiting platform for youth athletes. I build AI features behind evals, and I’ve killed work that didn’t earn its keep.
Experience
AppFolio · Product Manager, Rental Applications
- Sole PM for the application line: strategy and roadmap across applicant and property-manager experiences, three stakeholder types, two codebases that have to move in concert. The payer and the user are different people; the product has to serve both.
- Rebuilt the rental application end to end. Submission rates climbed meaningfully; properties on the new flow fill vacancies 3–5 days faster than offline workflows.
- Shipped follow-up for incomplete applications inside the authenticated app, not email or text, keeping every interaction auditable for screening compliance; AI review to trigger those requests automatically is next.
- Led the Zillow prefill partnership to 70% customer enablement, then made the case to wind it down when the data showed removing friction lowered applicant quality. Some friction is good friction.
- Run 50–100 direct customer calls a year and synthesize them with AI into a ranked pipeline of discovery endpoints.
- Pioneering spec-driven agentic engineering with my team using GitHub’s Spec Kit: object models feed the specs, behavioral specs and evals gate each release.
Lively · Data Scientist → Product Manager
- Built partner APIs and migration pipelines that moved $50M+ in HSA assets across financial institutions with zero data loss.
- Authored the data model and integration patterns partners used to launch a compliant HSA offering without building their own infrastructure.
- Stood up the company’s first data-science function; promoted to PM in 2021 as the work became the roadmap.
Oracle · Data Scientist, Applied Research
- Built an automated model-evaluation system, hundreds of tests per model update, gating releases on statistical significance rather than gut feel.
- Modeled large-scale consumer location and behavioral signals to predict purchasing.
Skills
Agentic AI product · eval design · JTBD · OOUX · experimentation · SQL · Python · APIs & data pipelines · LLM systems (RAG, tool use)
Selected Work
Brava (bravamade.com) — recruiting profiles for women’s soccer athletes, built with my brother. Coach-verified stats, benchmark rankings, an AI-assisted production pipeline. Designed and shipped end to end: research, product, brand.
AbiertoClaw (github.com/be-ns/AbiertoClaw) — open-source AI assistant over iMessage, built on parental leave to run our family calendar without touching the baby-leave budget. Swappable model sources, a complexity router, and a daily best-free-model picker keep it free by default.
Education
B.A. Communication Studies, Westmont College · Data Science Immersive (ML), Galvanize