Human-reviewed AI workflow
AI Inbox Triage & Drafting Assistant
Developed as a Johns Hopkins academic project, this inbox-triage system classifies, summarizes, and drafts safe first responses—then evaluates its own outputs against the source emails.
Practice
AI & Media Operations
I supervise metadata specialists at Paramount and work hands-on in streaming quality control for Disney, bringing real production constraints to applied AI.
What I do
I work at both levels of media operations: supervising specialists, priorities, and escalations while staying close to the assets and decisions themselves. That combination shapes how I design AI-assisted workflows: they need to help the people doing the work, produce results that can be reviewed, and leave important decisions with humans.
- Operations leadership: priority setting, escalations, training, quality metrics, tool testing, and workflow improvement
- Streaming QC and localization: video, multitrack audio, timed text, metadata, delivery troubleshooting, and Mandarin-language APAC content
- Applied AI: Johns Hopkins work in prompt design, retrieval, evaluation, and human-reviewed decision support
Selected work
Independent portfolio project
Media QC Guidance Copilot
A grounded-assistant design study built around retrieval before generation, visible sources, explicit uncertainty, fail-closed behavior, and accountable human review.
How I approach AI
I start with the operational problem: where assistance would help, what a good result looks like, and what should remain under human control. My applied AI studies at Johns Hopkins have strengthened my focus on clear context, evidence-based evaluation, privacy, and the consequences of getting things wrong.
I use OpenClaw to test these ideas in working assistants and prototypes. The aim is not automation for its own sake, but tools that help people retrieve, compare, draft, and think while keeping uncertainty visible and important decisions accountable to humans.