Steven Yu Email and Phone Number
Steven Yu work email
- Valid
- Valid
Steven Yu personal email
Steven Yu is a Deep learning and GPU Acceleration at Aurora. Performance optimization & Inference Optimization at Aurora. He possess expertise in java, c++, android development, opencv, management and 18 more skills. He is proficient in English.
Aurora
View-
Staff Software EngineerAurora Aug 2024 - PresentMountain View, California, UsEnjoying my day-to-day work to help Aurora ship the future! -
Senior Software EngineerAurora Nov 2022 - PresentMountain View, California, UsDeep Learning Acceleration TeamWorking on GPU, CUDA, pytorch, performance optimization, Inference Optimization and tooling -
Software Engineer, Performance OptimizationNvidia Apr 2020 - Oct 2022Santa Clara, Ca, UsPerformance Engineering and Optimization at Autonomous Vehicle E2E App Team. In charge of performance bringup for HW platform transition from Xavier to Orin SoC.Performance budget planning and high-level Performance KPI define, metrics collection, automation, regression analysis.Use various tools (perf, Nsight System, Nsight Compute etc.) to identify perf bottleneck, working with various teams(algorithm, OS, compiler...) together towards perf improvement of 50% E2E latency reduction in a 1.5 year span.Some of my work:* Cuda graph develop and integrate into DriveWorks DNN framework* Perception per/post processing CUDA Kernel optimizations for TensorRT/DLA (mem bound) * Motionplanner CUDA Kernel optimizations (compute bound)* Cloud Large-scale dataset replay, optimizing GPU Cost by boost GPU utilization by 50%. -
System Software Engineer- Autonomous VehiclesNvidia Aug 2018 - Mar 2020Santa Clara, Ca, UsSystem Software Engineer at Driveworks SDK Team, provide automative-grade middleware with high performant modular functionalities to developers. Based on NVIDIA Xaiver Tegra SoC.Compute Graph Framework and Heterogeneous Deterministic Scheduler* Developed NvsciBuf, a KMD utilizing NvRm NVIDIA GPU kernel driver api. Shipped in official CUDA Programming API. https://developer.nvidia.com/blog/sharing-cuda-resources-through-interoperability-with-nvscibuf-and-nvscisync/ * Developed cross-process channel between nodes connection in graph. Zero-cpy Exchange IPC utilizing nvstreams.Sensor Abstraction Layer * Camera Module API development, reduced offline video replay h254 decoding latency * Camera server to stream images across processes, cross-process nvmedia-cuda interop by eglstream -
Software EngineerWondergate Augmented Reality May 2017 - Aug 2017Developed build system for Augmented Reality App. Cross compile C++ code to Windows, Android, IOS based platform simultaneously
-
Research InternNational Laboratory Of Pattern Recognition, Chinese Academy Of Sciences Jul 2015 - Jul 2016Haidian, Beijing, CnDeveloped an DNN based image recognition algorithm by Caffe. Shipped algorithm to production on Android, by JNI and NDK -
Champion TeamHackshanghai Nov 2014 - Nov 2014Shanghai, CnCreated a cooking simulation game called "Leap Kitchen" based on Leap Motion Controller.
Steven Yu Skills
Steven Yu Education Details
-
Carnegie Mellon UniversityComputer Science -
UdacityComputer Science -
Nanjing University Of Information Science And TechnologyInformation System Engineering -
Peking UniversityComputer Science
Frequently Asked Questions about Steven Yu
What company does Steven Yu work for?
Steven Yu works for Aurora
What is Steven Yu's role at the current company?
Steven Yu's current role is Deep learning and GPU Acceleration at Aurora. Performance optimization & Inference Optimization.
What is Steven Yu's email address?
Steven Yu's email address is hs****@****dia.com
What schools did Steven Yu attend?
Steven Yu attended Carnegie Mellon University, Udacity, Nanjing University Of Information Science And Technology, Peking University.
What skills is Steven Yu known for?
Steven Yu has skills like Java, C++, Android Development, Opencv, Management, Public Speaking, Research, Computer Vision, Python, Deep Learning, Matlab, Mysql.
Free Chrome Extension
Find emails, phones & company data instantly
Download 750 million emails and 100 million phone numbers
Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.
Start your free trial