tapsa.ai ↗
A Wikipedia-sourced knowledge graph you explore along a night-sky timeline. Built end to end and shipped on my own, then debugged the real post-launch failures on nights and weekends.
Claude APICursorVercel
Director of Product · Workday
— Ships AI on nights & weekends
I build products to help organizations transform their workforce in real time.
Director of Product Management at Workday, running a nine-figure Workforce Planning line for 1,000+ enterprises. Off the clock I ship AI products solo and teach the next wave of PMs to build, not just theorize.
NowDirector of PM, Workday — Workforce Planning
Buildstapsa.ai · kiness.ai · micasita.ai
TeachesAI-native PM cohort · multi-university
FIG. 01 — B. HARO
01
Three sides of one problem: work.
Build
Build for transformation in real time
At Workday I own the Workforce Planning line — nine-figure ARR, 1,000+ enterprise customers, a team of senior and staff PMs. The job isn't shipping planning tools. It's building for companies whose shape is changing while they're using them.
Navigate
Sit with the change
I work alongside customers in the middle of a massive workforce transformation — rethinking headcount, skills, and roles that didn't exist a year ago. The planning problem just got harder, and far more interesting.
Prepare
Ready the people
I co-teach an AI-native PM cohort alongside a university faculty member — students from several schools getting genuinely hands-on with AI tools, so they enter the workforce able to build.
02
Why I stay in the weeds.
You can't develop intuition for a technology by reading about it.
Intuition comes from use — building with AI daily, pushing it until it breaks, and fixing what broke. That's what makes the pattern visible: where a company's work is actually about to change, which problems are now worth solving, and which opportunities nobody has named yet.
You can't tell an enterprise how AI will reshape their workforce if you've only ever seen it in a deck. So I stay hands-on — not as a hobby, but because it's the input to the judgment I'm paid for.
03
Things I've built, shipped, and debugged solo.
kiness.ai ↗
A dashboard for eldercare facilities that keeps families connected to what's happening with the people they love — the kind of unglamorous problem software should actually solve.
Full-stackLLMSQL
micasita.ai ↗
One calm home for everything a family is juggling. Calendar, to-dos, groceries, and budget in a single shared space — so the mental load stops living in one person's head.
Solo buildShared household
Weekend experiments
A running stack of small builds and prototypes to stay current with what's actually shipping in AI — the surest way I know to keep product judgment honest.
Claude CodeMCPLLM evals
// Small builds, real users, real failures. Nothing sharpens judgment faster.
04
Track record.
2021 — Now
Director of Product Management · Workday
Senior Manager → Director. Own the Workforce Planning line — nine-figure ARR, 1,000+ enterprise customers, a team of senior and staff PMs. More than doubled the business over three years.
2025 — Now
Solo builder & instructor · AI products · AI-native PM cohort
Built and shipped AI products solo, co-taught an AI-native PM cohort alongside a university faculty member, and advise an early-stage enterprise AI team. Hands-on daily with the Claude API, Claude Code, MCP, and LLM evals.
2021
Senior Product Manager · Human Interest
Owned the largest self-serve 401(k) product in a compliance-heavy environment. Raised participation to ~70%, grew AUM to ~$2B, and built the A/B testing framework from scratch.
2020 — 2021
Product Manager · Roofstock
Owned seller self-service for a real-estate marketplace. Launched a listing flow that grew active supply ~30% and cut cycle time ~20%.
2015 — 2020
APM → Senior PM · Workday
Shipped Workday's first AI-driven ERP automation — machine-learning anomaly detection across millions of financial journal lines, cutting enterprise month-end close time ~30%.
Education — 2022MBA · UC Berkeley, Haas
Education — 2013BA, International Relations · CSU Chico
Off the clock
Competitive marathoner — 2:46 PB, Boston qualifier. Running keeps effort and outcome honestly coupled. I want products to behave the same way.