I am an engineer and startup founder based in Silicon Valley. In my past life, I was a tank platoon commander in the Singapore Armed Forces and also led recruitment for Shopback (a Series E startup in Singapore) before I arrived in the US and earned my B.Eng in Computer Science from Cornell University.
I enjoy leading teams to tackle highly ambitious projects, and have worked in wildly different domains from the military, to ecommerce, to biotech and now in self-driving vehicles.
As an engineer, I consider myself as a full-stack across web frontend, backend, infrastructure, data and ML. I have an especially strong background in data processing and management for Machine Learning and scientific research, having done this in some really challenging domains (biotech and self-driving vehicles).
Places I have worked at and projects I have worked on include:
Eventual (YCombinator W22)
I Co-founded Eventual, which is building the data platform for binary data.
We are building our data engine Daft as an open-sourced product. It is a data engine like Spark, with a Dataframe frontend API, but highly tuned for workloads on binary type data such as images, video audio and more.
Learn more at www.getdaft.io.
Lyft L5 (Acquired by Woven Planet)
At Lyft L5, I was part of the inaugral L5 ML platform team that was responsible for building the data and compute tooling that powered all ML efforts at L5.
- WICKER: High throughput data processing, storage, analysis and serving for large-scale distributed deep learning (open-sourced)
- JADOO: Large scale data-parallel distributed training of neural networks
- GENERALIZED SENSOR FUSION: Flexible neural network architectures for multimodal sensor inputs (paper presented at Neurips 2021)
- SUPERVISED METRICS: Framework to measure and triage regressions in deployments of models on real world data
At Freenome, I pioneered and built the Research Platform team with 2 other founding engineers. Together, we built from scratch some of the foundational systems that continue to power Freenome’s cancer research platform today.
- CALZONE: Data warehousing for clinical and laboratory data
- FEATUREIO: High throughput data cataloguing and serving for ML experiments on multiomic (proteins, mass-spec, next-gen sequencing) data
- PINEAPPLE: YAML-based experiment and reporting framework for running large (100,000+) ML hyperparameter sweeps
- DAIKON: JupyterHub and Dask-based platform for interactive experimentation on Kubernetes, migrated researchers away from SLURM for up to 75% cost savings
- LIMS: Web application to manage complex wet-lab workflows for hundreds of thousands of blood samples
At Shopback, I led recruitment and HR. I doubled the company size and built teams from scratch in Singapore, Malaysia and Indonesia.
- Headhunted and hired key senior positions, including the Country GMs in Indonesia and Malaysia
- Hired for technical positions across all of software engineering, including their mobile team lead
- Aide to the CEO for fundraising and hiring across the Southeast Asia region
Singapore Armed Forces, Tank Battalion 48SAR
I was a Lieutenant of the Singapore Army - Platoon Commander, 3rd Company, 2nd Platoon.
- Commanded a platoon of tank crews in large scale military live-firing and operational exercises
- Topped my class in officer cadet school and underwent training for tank company command in Singapore’s military reserve
At Cornell, I was involved in computational biology research, on-campus startups and venture capital.
- COO, Cornell Venture Capital: Ran a student organization at Cornell that worked with some of the top Venture Capital firms in the US (NEA, Google Ventures, Sequoia, Bessemer etc)
- Research in the Yu lab on deep learning methods for protein folding
- Research in the Danko lab on interpretable deep learning methods for understanding Pro-SEQ data