Zhang
Leading ML Training in Uber | We are Hiring !!
- Role
- Engineering Manager - Uber Ai Platform (Michelangelo) at Uber
- Location
- Redmond, WA, US
- LinkedIn followers
- 500 followers
About Zhang
Engineer Manager of Training team in Uber AI platform (Michelangelo)- Expert in Deep Learning, distributed GPU training and serving, Generative AI- 13+ years experience in Machine Learning and Software Development- Proficient with PyTorch/Lightning, TensorFlow/Keras, Horovod(contributor), Ray, Huggingface, SparkML, Pandas, numpy, sklearn, deepspeed.
Experience
Engineering Manager - Uber Ai Platform (Michelangelo)
Aug 2018 — Present · Seattle, WA, US
2020 - Now, Leading the ML training team in Uber\'s AI platform (Michelangelo), responsible for scalable infrastructure supporting all ML model training and batch inference needs across the company. ML Training Framework: Built a unified distributed training service supporting cases like ETA, LLM, recommendation, etc. Scale training jobs across multi-GPU clusters, efficient data loading with strong reproducibility.Delivered interactive environments for experimentation and production pipeline workflows.Supported Tensorflow and PyTorch with data and model parallelism (Horovod, DDP, DeepSpeed), and advanced features like transfer learning and ensemble training.ML Infrastructure: Managed Uber’s ML resources across cloud providers and on-premise datacenters.Designed dynamic job schedulers for Spark and Ray workloads across multi-region clusters.Led resource planning, cost attribution, and GPU utilization optimization for production teams.Before 2020, TL for modeling Owns pipeline for Spark stream processing and machine learning modeling using phone sensor data. Tech lead for machine learning models for driver activity detection, offensive driving identification, fraud detection, ETA estimation to improve uber safety, map accuracy, and support ticket cost.
Skills
- Natural Language Processing
- Machine Learning
- C#
- Vhdl
- Simulations
- Java
- Testing
- Big Data
- Internet of Things
- Apache Spark
- Lte
- Azure
- Python
- Telecommunications
- Embedded Systems
- Deep Learning
- Matlab
- Object Oriented Design
- Antennas
- Algorithms
- Wireless
- Apache Kafka
- Neural Networks
- C++
- Wireless Networking
- C
- Power Bi
- Time Series Analysis
- Data Mining
- Digital Signal Processors
- Cloud Computing
- Linux
- Stream Processing
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.