Career update: After closing a ~5-year chapter working at Twitter, X, and xAI, I have joined @architectlabs 🚀
@xAI (now SpaceX) is truly special. It is where everything begins. The recos team taught me how to execute well, while the Timeline / Product team trained me how to ship fast with quality as a leader. Working with Elon over the years taught me "The Elon Algorithm," and the sense of urgency will stay with me. Ad astra to the friends since 2023 who went through the trenches with me and achieved incredible things under insane deadlines. Maybe the real AGI was the friends we made along the way.
Born and raised in Taiwan, educated at NTU EE (I really miss Prof. Lin-Shan Lee's "信號與人生" class), and with 99.9% of my friends and family in the EE industry across TSMC/MediaTek/(and many!), I know a chip design revolution is coming—and it excites me that I get to be a part of it.
Now, I’m excited to take the leap and build something new with truly talented leadership @axi_master and @aadityasubedi_. Architect Labs gathered a team of top-tier hardware industry veterans who know exactly where the real-world pain points are. Combined with a frontier AI lab spirit, we are working relentlessly to solve these issues. To me, this unique mix of talent is a strong signal that this is where the magic is going to happen.
In @architectlabs, we're building the AI that designs silicon, for your specifications. We’ve raised a $24M seed round (led by @stevejang@KindredVentures), and we are just getting started.
I’m happy to share that I’m starting a new position as Software Engineer at @anyscalecompute ! Thanks for @hsun_kai 's help. I will continue contributing to open-source @raydistributed project to make it even better!
The second article in the Kubernetes Operator tutorial series introduces how to write an operator using client-go, which is a relatively low-level approach. Higher-level frameworks will be introduced in later articles.
https://t.co/nyCFcGQcp6
I've been working on connecting GraphRAG to a local Llama model. It turned out to be more challenging than expected, requiring a monkey patch or an embedding model proxy server, so I’m documenting the monkey patch process.
https://t.co/g1kFgBCr2P
I'm starting a series on Kubernetes Operators! This first post introduces Custom Resources and Operators. Future posts will cover implementation details and compare different libraries.
https://t.co/qgmSCAJj8I
(1/n)
Training LLMs can be hindered by out-of-memory, scaling batch size, and seq length. Add one line to boost multi-GPU training throughput by 20% and reduce memory usage by 60%. Introducing Liger-Kernel: Efficient Triton Kernels for LLM Training.
https://t.co/nREx5sytwG
I used to think writing articles was time-consuming, but I found it sometimes saves my time. Lately, I've been mentoring some people on open-source, so I'm writing more tutorials and ask them to read everything before coming back with questions. 😂
https://t.co/mzTbS0py64
Whenever I’m in meetings and people see the tools I’m using, they usually ask what they are, and some even suggest me to start an online course 😂. So this time, I’m writing the introduction of one tool I recommend the most for developers.
https://t.co/y9Lr1VMZLa
Recently, some junior engineers expressed interest in developing open-source projects. To avoid having to teach each newcomer the process repeatedly, I took a few minutes to write a super beginner level guide on the Git workflow for open-source projects.
https://t.co/U8geu26Pow
I was very glad to attend @ApacheCon as a volunteer last month. I wrote a blog post to share my volunteer experience and takeaways. Specially thanks for the CommunityOverCode team and the Apache Travel Assistance Committee!
#apache#CommunityOverCode
https://t.co/W577qibzjF
New article is here! Can you believe my personal website is completely free except for the domain name? This time, I discuss the important details to pay attention to when deploying a personal website.
https://t.co/IrZXKKPoJt
I decided to start writing technical blogs! Recently, I revamped my personal website, and this is my first article, which discusses how to simulate Kubernetes' Node-pressure Eviction in a local environment.
https://t.co/Vo3ax1a0gn