Mengyan
(Derrick) Yao

I find a real systems problem, reverse the workflow, and ship a production platform.

I build data automation, applied AI workflows, and digital systems that turn fragmented business operations into scalable infrastructure.

Now: Data Automation & Digital Systems Manager / Lead iHOMI (International House of Music) · Los Angeles · Jun 2025–Present
Based: Monterey Park / Los Angeles US Permanent Resident

Selected work

90%

less manual work · weeks → min

iHOMI · Los Angeles · 2025–Present

DaOS Retail

Catalog work at International House of Music ran across vendor files, SKU matching, pricing, AIMsi ERP/POS, and Magento. I built the internal automation platform that now does that work. Update cycles collapsed from weeks to minutes; manual effort dropped by about 90%. DaOS Retail classifies and audits catalog taxonomy with an LLM, publishes from supplier data and manufacturer pages, and puts orders, exceptions, warehouse, and CRM on one operations dashboard.

Rust-native

supplier pricing DB

DaOS Retail · iHOMI

DotB

A Rust-native supplier pricing database and search engine under DaOS Retail. Vendor price sheets become a fast, searchable, source-verified system, with incremental updates and indexing built for that job.

Live IR

malicious PHP payload

iHOMI · incident

Production incident

During a live cyberattack I reverse-engineered a malicious PHP payload, traced attacker infrastructure, and escalated forensic findings so the team could contain it.

PMS API

hospitality ops

Zoey’s Apartments / Nuanming Xiyue · 2020–Present

DaOS Hospitality

Co-founded the brand. Built DaOS Hospitality as the operations layer: PMS API integration, competitor price collection, and dynamic pricing.

Experience

Jun 2025–Present

Data Automation & Digital Systems Manager / Lead

iHOMI (International House of Music) · Los Angeles

Own the stack that connects vendor data, SKU matching, pricing, AIMsi ERP/POS, Magento, warehouse, and CRM. Ship internal platforms so operations can run without waiting on engineering for every catalog or order exception.

Jan 2024–Apr 2024

AI Trainer

Outlier / Scale AI · Remote

Evaluated 500+ model outputs. Wrote domain-specific queries and structured feedback for model evaluation.

Sep 2022–Jan 2023

Research Assistant

University of California, Santa Barbara

Analyzed 50+ K–12 education datasets to surface rollout patterns and reporting bottlenecks. Cut reporting time by 25%.

May 2021–Jul 2021

Analyst Assistant Intern

Capgemini China · Shanghai

Python and Tableau supply-chain dashboards for a Fortune 500 client. Saved 15+ hours a month of tracking and executive reporting.

2020–Present

Co-founder

Zoey’s Apartments / Nuanming Xiyue

Hospitality brand operations layer — DaOS Hospitality — covering PMS API integration, competitor price collection, and dynamic pricing.

Research

Independent systems work: firmware and userland reconstruction, responsible security research, and applied computer vision. Tesla pre-approved security researcher.

overlayfs / QtCar

on WSL

Infotainment · registered research vehicle

Windows Subsystem for Tesla

Brought Tesla infotainment userland up outside the car on Windows/WSL — overlayfs runtime reconstruction and QtCar HMI. Identified a parse–authorize–serialize–reparse authorization bypass in Service Shell and validated it on a registered research vehicle. Disclosed as a Tesla pre-approved security researcher.

Persistent root

SSH / ADB

Embedded Linux

Insta360 Luna Ultra

Persistent root SSH and ADB on the Luna Ultra, with arbitrary userland execution — including DOOM, as a proof of control.

≤1 in

error · 6K+ samples

Computer vision

Body measurement

EfficientNet-B3 and KeypointRCNN predicting body measurements from images. Combined 6K+ samples; ≤1-inch error.

Education

Sep 2024–Dec 2027

M.S. Computer Science

Georgia Institute of Technology · OMSCS

Jan 2023–Aug 2024

M.S. Analytics

University of Southern California

Sep 2020–Dec 2022

B.A. Statistics & Data Science; Education minor