Develop HPC systems for RDC.

Storage system trouble shooting.(JBOD Lustre Gluster Hadoop VROC etc)

GPU system performance testing.(HPL, TensorRT,etc).

• Senior Linux administrator

Maintain the automatic sequencing IT system at CGI and transfer them to BGI.

Installation with DHCP+Pxe+Kickstart.

Configuration with puppet.

Monitoring with Ganglia and nagious system.

Remote control with Ipmitool (remote screen redirect) and Megcli (command line make raid) Jobs schedule software Scrat (writen by CGI) and SGE to manage the jobs.

Deploy and maintain shiny-server (rstudio) for research and develop team. Maintain a Git server.

Bash script for crontab and monitor.

Help research team develop a new sequencer’s IT system with 100TB/run level throughput, and a small desktop sequencer. Including the following technologies:

Mellanox infiniband FDR 56Gb/s, Big mtu (65000), bonding, partition, IPoIB, RDMA.

Nvidia Cuda 7.5 environment in Centos 7.

4 way 4TB memory server.Hugepage memory as a local disk.

Mapr system for storage, Isilon storage, Huawei storage, gcluster storage.

Focus on high-speed storage and achievements of high throughput DNA sequencing data, consultation for China gene bank with a 50PB tier storage platform (iRods and archive solution).

Senior system engineer: design, built, deploy, maintain and update the HPC cluster to make the HPC platform easy to deploy, maintain and scale suitable for different Labs. The platform with 22,036 CPU cores and 20.6PB storage can support thousands of bioinformaticians analyzing more than 10TB data per day.

Data Center builder: I am one of the four main founders of High Performance Computer (HPC) platform at BGI which is the world’s biggest genomic sequencing company. We started from scratch and established all of six BGI Data-Centers in China, Europe and America.