Lai Wei
Inference @ Meta Superintelligence Lab, ex-AWS AI
About
I'm an Engineering Manager at Meta Superintelligence Lab working on LLM inference engines.
My inspiration comes from Person of Interest (2011) — a show where an advanced AI called the Machine is trained by its creator using something that looked a lot like RLHF before RLHF had a name. It's what got me interested in AGI.
Throughout my career I've been lucky to work on cutting-edge projects with talented teams who share that vision.
Projects
Proof of Concept: AI Misuse
Submission to OpenAI's Preparedness Challenge.
Writing
Read my blog →- Tips and tricks building custom GPTs
- Build high-performance ML models using PyTorch 2.0 on AWS
- LG AI Research develops a foundation model using Amazon SageMaker
- Amazon SageMaker 大规模机器学习实践
- AWS and NVIDIA achieve the fastest training times for Mask R-CNN and T5-3B
- How Netflix uses DJL for distributed deep learning inference in real-time
- Detecting pneumonia from chest x-ray images in Java
- Using DJL to deploy a deep learning model from Apache Kafka
- Apache MXNet 1.5.0 release
- Apache MXNet (incubating) adds support for Keras 2
- Train using Keras-MXNet and inference using MXNet Scala API
- MXNet 开放支持 Keras，高效实现 CNN 与 RNN 的分布式训练
- Introduction to deep learning for developers· slides
- RecSys 2017: Transfer learning for personalized content and ad recommendation