ML Engineer · GPU Hacker · Builder

Jack Min Ong

Machine Learning Engineer at PrimeIntellect, working on distributed training infrastructure, LLM inference serving, and GPU kernel development. I think a lot about making models train faster and serve cheaper — from CUDA kernels on Blackwell GPUs to orchestrating clusters at scale.

Born in Penang, Malaysia. Based in San Francisco.

Building in the open SF / PG / KL

Find Me

Writing

All posts →

Kitchen Sink

GPU Cluster Utilization
tok/s Throughput (live)
World Clock
--:--
SF
--:--
Penang
--:--
London
Interests Orbit
Commit Activity
// dev/null
System Status
vLLM cluster
K8s nodes
NVLink mesh
RDMA fabric
Uptime: 0d 0h 0m
Mem: -- / 141 GB HBM3e