avatar

Siyoung Oh

Zen and the Art of System Engineering

I’m a software engineer at Snowflake. I’m interested in building simple but powerful large-scale systems that can unlock diverse computing capabilities at a low cost. While people have gotten used to the outstanding capabilities of computers, what computing can enable upper layers to achieve, like AlphaGo or ChatGPT, is still truly mind-blowing. Yet, we have a lot to do to up-level the underlying layers to unleash human potential, and I’m excited to be a part of the wave.

In a larger context, I’m interested in enhancing the System (e.g., human health, company values, or cluster reliability and efficiency).

Past 🔗

Previously, I worked on Google’s cluster management system (Borg) for six years and Quora’s ML platform for two years. In the Borg team, my last role was a TLM to transition homogeneous cluster abstractions to heterogeneous machine families while enhancing reliability and efficiency. In Quora, I streamlined the ML development lifecycle from experiments to serving to easily adopt diverse options (e.g., TF, PyTorch, Transformers) and boost productivity with more principled container engineering.

Before that, I worked in Korean companies (Kakao, Nexon), participated in competitions (TopCoder, ICPC, IOI, Putnam), interned in three different companies (Google, DRW Trading, Facebook), and studied mathematics and computer science at CMU. As I navigated the journey, I learned that a few career decisions are as vital as hard work, so I wrote a few about the topic (link: career ). ∎