CM3 Consulting
Founder
Scott Ziolko
Data science leader experienced in AI/ML, business strategy, neuroscience, consumer goods, ad tech, market research, logistics, fintech & education.
Data science leader experienced in AI/ML, business strategy, neuroscience, consumer goods, ad tech, market research, logistics, fintech & education.
There was a problem loading your content.
Experience
Carnegie Mellon University
Employee
University of Pittsburgh School of Medicine
Employee
Geoscape
Employee
Latin Force Group LLC New York City
Employee
Broward College
Employee
Lead Data Scientist / Head of Data
Lead an AI-native data function at a business banking fintech, as a hands-on leader, using Claude, MCP servers, and agents to deliver company-wide impact across risk, finance,... more operations, and sales.
• Self-Service Analytics: Integrated Claude with custom MCP servers for Snowflake, Metabase, dbt, and accounting tools, plus dbt documentation published into Notion. The executive team plus product, compliance, underwriting, and engineering now get answers from Claude directly while I lead expert review of AI-generated analysis on critical company projects; self-service handled all new dashboard requests in 2026 with data review on critical reports.
• Automated Underwriting: Replaced a manual, spreadsheet-driven credit review with an automated system, cutting turnaround from ~1 business day to under 60 seconds. It evaluates 100% of credit requests (tens of thousands to date), ~10% as the primary decision and the rest in shadow mode. Built on Snowflake, dbt, and tested shared Python libraries, with a Streamlit app for human-in-the-loop review and full audit trails.
• Credit Risk Modeling: Improved AUC by more than 0.15, using only information available at sign-up so no future data could leak into training and contaminate the results. Back-testing showed it would have roughly doubled credit for the best customers while eliminating more than half of historical bad debt; validated in a 6-month champion/challenger shadow test.
• Profitability Metrics: Identified which customer segments and behaviors drove margin, guiding priorities as the company moved from negative margin to a positive monthly P&L.
• Agentic Workflow Automation: Built a role-restricted QuickBooks MCP with SSO so Claude books new customers and vendors and answers cash- and accrual-basis questions from live accounting data; added AI coaching that grades every account manager call against a rubric, and a cached OpenAI GPT classifier for new transaction types.
• Data Engineering Standards: Used Claude as the primary coding agent with custom skills encoding style, testing, and best practices; restructured dbt with SCD Type 2 snapshots, linting, automated tests, and GitHub Actions CI/CD.
• Self-Service Analytics: Integrated Claude with custom MCP servers for Snowflake, Metabase, dbt, and accounting tools, plus dbt documentation published into Notion. The executive team plus product, compliance, underwriting, and engineering now get answers from Claude directly while I lead expert review of AI-generated analysis on critical company projects; self-service handled all new dashboard requests in 2026 with data review on critical reports.
• Automated Underwriting: Replaced a manual, spreadsheet-driven credit review with an automated system, cutting turnaround from ~1 business day to under 60 seconds. It evaluates 100% of credit requests (tens of thousands to date), ~10% as the primary decision and the rest in shadow mode. Built on Snowflake, dbt, and tested shared Python libraries, with a Streamlit app for human-in-the-loop review and full audit trails.
• Credit Risk Modeling: Improved AUC by more than 0.15, using only information available at sign-up so no future data could leak into training and contaminate the results. Back-testing showed it would have roughly doubled credit for the best customers while eliminating more than half of historical bad debt; validated in a 6-month champion/challenger shadow test.
• Profitability Metrics: Identified which customer segments and behaviors drove margin, guiding priorities as the company moved from negative margin to a positive monthly P&L.
• Agentic Workflow Automation: Built a role-restricted QuickBooks MCP with SSO so Claude books new customers and vendors and answers cash- and accrual-basis questions from live accounting data; added AI coaching that grades every account manager call against a rubric, and a cached OpenAI GPT classifier for new transaction types.
• Data Engineering Standards: Used Claude as the primary coding agent with custom skills encoding style, testing, and best practices; restructured dbt with SCD Type 2 snapshots, linting, automated tests, and GitHub Actions CI/CD.
Head of Data
Led data science, data engineering, and analytics for a consumer TV streaming service on mobile, desktop, and smart-TV platforms as it grew from 300K to 1.5M subscribers (5x);... more built a team of 10+ data scientists, analysts, and engineers, including two managers.
• Experimentation Platform: Architected a self-hosted experimentation platform (GrowthBook on Segment and dbt modeled event data) and set standards for A/B, holdout, difference-in-differences, synthetic control, media mix modeling, and incrementality testing. Experiment volume grew exponentially for 4 consecutive years across 6 product teams.
• North Star Metrics: Defined active subscribers and sessions reaching at least one ad pod as company north stars. Redesigned conversion funnels and lifetime value when a new free tier changed how user value had to be measured.
• Pricing Research: Designed and ran conjoint surveys to segment users and set pricing strategy, leading to new price tiers and price increases that improved margin and contributed to the first GAAP-profitable year.
• System Performance Analysis: Analyzed petabyte-scale logs of user and system behavior to find the platforms and services in the critical path for the most impactful performance improvements, focusing work on playback errors, time to first playback, and UI layouts.
• Executive Partnership: Partnered directly with the CEO and CFO on initiatives producing multi-million-dollar annual savings, critical to the first GAAP-profitable year; improved customer acquisition efficiency by more than 50%.
• Data Platform & Team Leadership: Scaled pipelines to trillions of events per month; hired and mentored PhDs in physics, biology, biostatistics, marketing, and computer science; prototyped LLM analysis of customer interactions.
• Experimentation Platform: Architected a self-hosted experimentation platform (GrowthBook on Segment and dbt modeled event data) and set standards for A/B, holdout, difference-in-differences, synthetic control, media mix modeling, and incrementality testing. Experiment volume grew exponentially for 4 consecutive years across 6 product teams.
• North Star Metrics: Defined active subscribers and sessions reaching at least one ad pod as company north stars. Redesigned conversion funnels and lifetime value when a new free tier changed how user value had to be measured.
• Pricing Research: Designed and ran conjoint surveys to segment users and set pricing strategy, leading to new price tiers and price increases that improved margin and contributed to the first GAAP-profitable year.
• System Performance Analysis: Analyzed petabyte-scale logs of user and system behavior to find the platforms and services in the critical path for the most impactful performance improvements, focusing work on playback errors, time to first playback, and UI layouts.
• Executive Partnership: Partnered directly with the CEO and CFO on initiatives producing multi-million-dollar annual savings, critical to the first GAAP-profitable year; improved customer acquisition efficiency by more than 50%.
• Data Platform & Team Leadership: Scaled pipelines to trillions of events per month; hired and mentored PhDs in physics, biology, biostatistics, marketing, and computer science; prototyped LLM analysis of customer interactions.
Head of Data
Fino Consulting
Head of Data Science
Founder
Employee
Education
About
Achievements
Bringing profitability to startups.
Describe the most impressive thing you've done.
Skills
Locations