AI & Creative Technology Program Lead
Fifteen enterprise productions. 3 Agentic AI tooling product and shipped. 100+ creator program at 3M+ monthly views. A game at #9 on the Epic wishlist. Eleven projects, one person holding strategy, craft, and code together.
I work hands-on with multimodal AI, APIs, and code to prototype workflows, contribute PRs, test ideas with users, and turn real usage into product decisions.
Designed, built, and operated a human-in-the-loop evaluation product as sole engineer — 14,390 judgments across 105 model builds. Grading throughput rose from ~10 to ~26 builds a month, saving $200K in vendor cost.
Built a custom AI agent on YouTube Analytics and market data to repackage a 1,888-video, 413K-subscriber channel for U.S. discovery. First month: U.S. views +55.5%, total views +32%.
Sole-built the data pipeline and interview platform behind Dipper's user research — every account segmented, eligibility-checked, and privacy-flagged. Restored 3,760+ users to visibility and put a 16-person interview team on one workflow.
Connected company strategy, technical launches, creators, internal production, global distribution, measurement, contracts, and payouts through one content-driven marketing engine.
Carried creative concepts through production, technical dependencies, reviews, QA, partner coordination, and delivery of 500+ campaign assets.
Connected artists, technical directors, developers, vendors, assets, tools, reviews, and delivery milestones through one global production pipeline.
Turned ambitious game concepts into executable plans across roadmaps, asset production, engineering, QA, releases, marketing, and community—helping the title reach #9 on the Epic Games Store wishlist.
Connected content, localization vendors, reviewers, regional stakeholders, quality control, schedules, and delivery across a multilingual operation.
Turned fragmented auction and market data into an intelligence product for art-market and wealth-management users.
Synthesized player feedback, aligned priorities with technical constraints, and converted community needs into an actionable feature-development roadmap.
Built annotation, quality-control, and evaluation workflows to train computer-vision models across varied airport and runway environments.