HOME
Hello, my name is Tze-Yang, I’m a Research Scientist and Entrepreneur, currently working at Nokia Bell Labs. My work is primarily concerned with developing the interfaces and communication frameworks to connect machine learning agents in distributed systems. The greatest impact of my research is the application of traditional concepts in information theory to create the underlying formalism for how AI systems can learn to communicate effectively and efficiently.
Professional Career
Member of Technical Staff
Nokia Bell Labs
My work is primarily concerned with developing the interfaces and communication frameworks to connect machine learning agents in distributed systems. The greatest impact of my research is the application of traditional concepts in information theory to create the underlying formalism for how AI systems can learn to communicate effectively and efficiently.
CEO/Co-founder
I co-founded WAIveform, a startup company that focuses on using machine learning to optimize wireless video transmission. Our patented proprietary algorithm simultaneously optimizes video compression and transmission achieving <1ms video latency and no video dropout whatsoever.
What I’m good at
Programming languages
I have over 7 years of experience in Python and C++, including over 5 years developing machine learning models using Pytorch and Tensorflow. I also have extensive experience using GNURadio with software defined radios (SDR). I have also recently started learning Rust and found it to be much better for my mental health than C++.
Hardware
Unlike most machine learning researchers, the models I develop do not simply live in simulation. I have extensive experience deploying machine learning models using C++ and Docker on NVIDIA Jetson devices.
Entrepreneurship
As a research scientist, I have always been fascinated by the interaction between the commercial world and research and how pressing issues in society can be solved by state-of-the-art research. I started my company to improve the reliability of wireless video streaming for drones and VR devices. Throughout the startup experience, I not only worked with people across the industry, including Royal College of Art students to understand their challenges when using VR to create, but also entrepreneurs, such as the CEO of Creator Fund, Jamie Macfarlane, who gave me invaluable advice on how to start a company. I also took a course from the Imperial College Business School that taught me the basics of startup entrepreneurship, where I raised the initial funding to start the customer discovery process.
System Admin
During my PhD at the Information Processing and Communications Lab (IPC-Lab), I was the system admin of the group’s GPU cluster. I was the admin for 3 years, growing the cluster from the initial 8 GPUs to now 30+ GPUs. We use a dashboard for monitoring GPU utilization so that everyone in the lab can have access to GPUs, and it is built using PHP and Apache.