I don't just define products—I can validate technical approaches, write proof-of-concept code, and speak fluently with data scientists, ML engineers, and architects. I'm the product leader who can debate gradient boosting parameters and explain the business case to the CFO.
The Unicorn: Product strategy + hands-on technical execution
Organized by the business problems I solve
Build predictive models that forecast churn, conversion, lifetime value, and next-best actions
Real application: Built ML churn model at USAA that saved $12M by predicting dissatisfaction 60+ days before customers left
Build real-time monitoring systems that catch compliance issues and fraud before they become problems
Real application: Led development of Control Tower/Sentry360 at ServiceMac—compliance platform with 5.2x ROI monitoring millions of transactions daily
Build intelligent routing, automation, and resource allocation systems that do more with less
Real application: Deployed real-time NLP routing at LoanCare that lifted conversions 18pp and reduced handle time 22%
Build segmentation, recommendation, and targeting systems that deliver 1-to-1 experiences to millions
Real application: Co-designed AI segmentation engine at Amazon that created 87 micro-segments and generated $38M incremental revenue
Why technical depth matters in product leadership
When your data scientist says "the model isn't converging," I can look at the learning curves, check for data leakage, and suggest regularization approaches. I don't just nod and trust the experts—I validate their work.
Before committing engineering resources, I prototype solutions in Python. This de-risks projects and proves feasibility before anyone writes production code. I've saved teams months by discovering technical constraints early.
I can explain gradient boosting to a data scientist and explain the business case for ML to a CFO. This translation ability is rare—and it's why cross-functional teams actually ship products instead of arguing about priorities.
I've debugged models in production, dealt with data drift, handled concept drift, and explained to executives why accuracy dropped 10 points overnight. Experience with ML in production makes me better at designing systems that won't break.
My process for taking products from concept to production
Customer interviews, data analysis, competitive research. I find the real problem—not the symptom everyone's complaining about.
Prototype solutions, validate technical approach with data scientists, define success metrics. I write SQL and Python to test assumptions.
Agile sprints with data scientists, ML engineers, and developers. I review code, debug models, and make trade-off decisions in real-time.
Track KPIs, run experiments, iterate based on data. I build dashboards and monitor model performance to ensure continuous improvement.
Tools and technologies I work with daily
Most AI PMs talk strategy. I can also validate architectures, write prototypes, and debug models. Let's discuss your technical challenges.