I am a Product Data Scientist who turns ambiguous product questions and messy behavioral data into measurable decisions.
I combine product judgment, experimentation, analytics engineering, and applied AI. I care not only about how a system is built, but also what should be measured, why it matters, and whether the evidence is trustworthy enough to act on.
- 📊 Define product KPIs, funnels, conversion metrics, and decision guardrails
- 🔍 Diagnose metric changes through segmentation, behavioral analysis, and data-quality investigation
- 🧪 Design experiments and evaluation frameworks that connect evidence to product decisions
- 🛠️ Turn repeatable analytical judgment into governed, auditable data and AI systems
| Area | Capabilities |
|---|---|
| Product Analytics | KPI Design · Funnels · Conversion · Segmentation · Retention · Metric Diagnostics |
| Decision Science | Experimentation · Causal Thinking · Guardrail Metrics · Decision Frameworks |
| Data and Analytics | SQL · Python · pandas · SQLite · Semantic Layers · Data Contracts · Data Quality |
| Applied AI | RAG · LLM Workflows · Tool-Using Agents · Structured Extraction · Human-in-the-Loop |
| Reliability | Golden Evals · Regression Testing · Governed SQL · Traceability · Abstention |
Helps product and business teams answer metric questions with consistent definitions, visible data-quality warnings, and auditable evidence. The system supports funnel and conversion analysis while using a semantic layer, governed SQL, deterministic orchestration, and golden evaluations.
Helps renters navigate complex housing rules by turning fragmented legal information into structured, explainable guidance. The system combines hybrid RAG, LLM extraction, rule-based reasoning, geospatial data, and deterministic evaluation.
I start with the user problem and product decision, define the metrics and evidence required, and then build the smallest reliable analytical system that can support that decision. My goal is not just a polished output—it is a solution that is useful, auditable, reusable, and honest about uncertainty.