Portfolio

Selected Projects

Newest-to-oldest project highlights across AI, data science, and enterprise transformation work.

  1. 2026 - Present

    Analytics Studio Copilot

    Built a specialized AI copilot for marketing leaders to measure performance, run diagnostics, and optimize cross-channel budget allocation with confidence.

    • Designed an agent harness with three loop layers to plan analysis, call tools, and refine results from follow-up questions.
    • Implemented intent and multi-intent routing with taxonomy plus core language modeling to map requests to the right skills and APIs.
    • Integrated FastAPI and Python machine learning services on Microsoft Azure with secure cloud deployment and compliance-safe access patterns.
    • Enabled marketing leaders to evaluate channel saturation, cut wasted spend, and reallocate budget to higher performing channels.
    • Reduced deep-dive analysis time from about 10 working days to about 5 minutes and supported 9 to 12 percent revenue growth outcomes for ecommerce teams managing over 100 million in spend.
    Abstract diagram of an AI copilot orchestrating marketing budget decisions

  2. 2024 - 2025

    Creative Intelligence Platform

    Multi-agent AI system that analyzes images and videos at scale to decode what drives creative performance, transforming manual tagging into automated insight generation for marketing teams.

    • Deployed multi-agent architecture with foundation models to label and tag creative assets across social media, YouTube, and other marketing channels.
    • Scaled creative analysis from slow manual processing to high-throughput automated tagging covering company and competitor content.
    • Built foundational creative measurement layer revealing relationships between creative attributes and audience engagement.
    • Enabled marketing leaders to understand performance drivers and optimize creative strategy with data-backed insights.
    Abstract visualization of AI agents analyzing video and image content to extract creative insights

  3. 2025 - 2025

    Supply Chain Automation Engine

    Designed and deployed AI workflow automation for a supply chain business to remove repetitive operational work across document processing, data entry, and system updates.

    • Prototyped end-to-end flows in n8n, then migrated to a production-grade LlamaIndex code workflow architecture.
    • Integrated multiple LLM models with a custom OCR and pattern detection layer for industry-specific PDF reports.
    • Connected automation to Bexio, accounting platforms, and internal tools to cover order registration, purchase flow, delivery, and invoicing.
    • Implemented human-in-the-loop controls for operational oversight while keeping throughput high.
    • Reduced manual tasks by 90 percent and enabled 70 percent more delivery capacity without increasing headcount.
    Abstract workflow map for supply chain AI automation with OCR, LLM, and business system integrations

  4. 2020 - 2025

    Marketing Mix Modeling Platform

    First enterprise-scale marketing mix model at EA, uniting finance, marketing, and branding teams to optimize $500M+ annual budget allocation across five franchises, channels, and global regions with predictive revenue forecasting and executive-grade analytics.

    • Built end-to-end platform from data governance and standardization through statistical modeling and interactive dashboards for quarterly VP and C-level decision support.
    • Grew from founding individual contributor to team lead, scaling the solution across five studio franchises and diverse stakeholder groups.
    • Delivered predictive models guiding channel and regional budget allocation, with unified taxonomy and reliable analytics foundation.
    • Enabled cross-functional collaboration and data-driven marketing investment strategy at unprecedented company scale.
    Abstract visualization of marketing mix modeling with budget optimization and revenue forecasting

  5. 2019 - 2024

    Player Lifecycle Intelligence Platform

    Enterprise behavioral prediction system spanning all major EA franchises (sports titles and live service games), forecasting player churn, conversion, and engagement to power hyper-personalized campaigns and cross-functional optimization.

    • Led 5-year platform development from inception through production, delivering high-accuracy propensity models across diverse game types and yearly release cycles.
    • Built API and agent-accessible services enabling real-time segmentation, campaign targeting, and churn prevention at scale.
    • Collaborated with Sony and Microsoft on first-party platform integrations for cross-ecosystem insights.
    • Enabled measurement and A/B testing infrastructure used across departments, driving measurably higher campaign performance than baseline.
    Abstract visualization of player behavioral prediction and lifecycle segmentation across gaming platforms

  6. 2025 - 2026

    Player Value Forecasting Engine

    Built enterprise-scale ML platform predicting two-year player lifetime value across 100M+ users in EA's flagship sports franchises (FIFA/FC and Madden), enabling strategic player engagement and revenue optimization.

    • Delivered robust statistical models handling heavy-tailed LTV distributions and cross-franchise cycle predictions spanning multiple game releases.
    • Productionized on AWS with weekly automated recalculation pipeline and Snowflake integration for real-time analytics access.
    • Enabled data-driven engagement strategies through player-level LTV forecasts that look beyond single game cycles.
    • Scaled inference to process massive player populations while maintaining prediction accuracy and business SLAs.
    Abstract visualization of player lifetime value prediction across gaming franchises

  7. 2016 - 2019

    Clinical Cancer Diagnostics Platform

    Pioneered SOPHiA GENETICS' first RNA-seq gene fusion detection pipeline, delivering CE-IVD certified clinical-grade diagnostics for solid tumors and leukemia that enabled precision oncology for 50,000+ patients globally with state-of-the-art sensitivity and specificity.

    • Built end-to-end detection algorithms for chromosomal rearrangements and gene fusions from research through production, post-sale support, and regulatory certification.
    • Collaborated with hospitals, oncologists, and laboratory scientists worldwide to deliver five commercialized cancer diagnostic products meeting clinical-grade reliability standards.
    • Worked with quality and regulatory teams to achieve algorithmic transparency and CE-IVD compliance for global deployment.
    • Minimized false negatives on real-world messy clinical data, maintaining top-class accuracy critical for life-saving treatment decisions.
    Abstract visualization of RNA-seq gene fusion detection for precision cancer diagnostics