Yan Wang

AI/ML Infrastructure for Physical AI (World Foundation Model) at NVIDIA Cosmos Lab

An online community for mandarin-speaking pilots, flight instructors and aviation enthusiasts.

Leading a team that builds a scalable, flexible and reliable PyTorch Lightning based distributed training platform that offers high performance model training, evaluation, experimentation, analysis and tooling suite for debugability, observability and benchmarking.

Leading a team to build machine learning training performance tools and partner with ML engineers to scale training for Cruise AI models to improve cost efficiency and accelerate ML.

Cruise AI, Machine Learning Platform. PyTorch based Machine Learning platform, dataloader, training performance

Amazon Halo Health Technology, ML Platform

Machine Learning Algorithms team, AI-Research organization • Built internal machine learning benchmarking platform on top of Azure Batch (NodeJS, C#, AngularJS, SQLServer, Azure Batch). • Implemented random forest algorithm in product (C#, decision tree, machine learning basics).

Windows Fundamental Compatibility team, Windows & Devices organization • Development two releases of Windows App Certification Kit. (C++, C#, Windows App technologies, scheduling algorithms, etc.)

Windows Ecosystem team, Windows & Devices organization • WDS (Windows Deployment Services): C++, C#, multi-thread, network, PXE, unit test, etc. • VAMT (Volume Activation Management Tool): C#, SQL, unit test, activation technologies. • Kits Building Infrastructure and ADK Installer: C#, WIX, KML, building engineering.

Grading students' homework, answer questions.

The project is to gather the fully dimensional data for the entire Google display network, and use Bayesian Hypothesis to test the independency between the dimensions. The result is used to collect factorized data instead of the currently full dimensions. MapReduce and Bigtable are used in the project.

* Master Thesis: GeoSocialMap: Visualization of Geosocial Data * Publication: Kurant, Maciej, Minas Gjoka, Yan Wang, Zack W Almquist, Carter T Butts, and Athina Markopoulou. Coarse-Grained Topology Estimation via Graph Sampling. ACM SIGCOMM Workshop on Online Social Networks (WOSN) '12

Bachelor Thesis: Research on Chord-based Query System Simulation