PinnedAutonomous Vehicle: a Compound AI SystemAI took the spotlight in 2023 once again, largely due to the success of ChatGPT, which is built on top of a pre-trained foundation model…Jun 4Jun 4
PinnedPublished inTowards Data ScienceChallenges of Machine Learning Platforms for Autonomous VehiclesAutonomous vehicles (AV) are the future of mobility. Thanks to recent technology breakthroughs, they no longer only exist in sci-fi movies…Oct 6, 2022Oct 6, 2022
PinnedPublished inTowards Data ScienceA Brief History of Deep Learning FrameworksA survey of the deep learning frameworks and their future projectionDec 8, 2020Dec 8, 2020
The Future of MobilityFour years ago, when I joined Waymo, there were no autonomous vehicles operating on the streets of San Francisco. However, last year marked…Jan 17Jan 17
Published inLevel Up CodingUnderstanding Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep LearningReading Notes of Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep LearningAug 12, 2021Aug 12, 2021
Published inTowards Data ScienceUnderstanding AdaSum: Adaptive Summation Method for Scaling Distributed TrainingML Reading NotesMay 2, 2021May 2, 2021
Published inTowards Data ScienceHow Much Does The Google Pay Promotion Cost?The math behind the $21 referral bonusDec 18, 2020Dec 18, 2020
Published inLevel Up CodingHow I Failed the FLAG InterviewsSome thoughts about mid-career transitionAug 22, 20201Aug 22, 20201
A Tale of Two Software IndustriesElectronic Design Automation (EDA) software is a set of tools and libraries that help semiconductor hardware engineers to design and/or…Jun 10, 2020Jun 10, 2020
Published inApache MXNetFrom Three Hours to 25 Minutes: Our Journey of Optimizing Mask R-CNN Training Time Using Apache…Written by: Karan Jariwala, Jerry Zhang, Lin Yuan @ Amazon and Przemek Tredak @ NVIDIAApr 1, 2020Apr 1, 2020
