Diving Deeper into Machine Learning

Liqian Ma (MSAI ‘22) worked in China at tech company Alibaba Group when he witnessed the potential of machine learning (ML). 

Ma was a senior machine learning engineer at the time, and he used mobile-device sensor data to identify users so vision-impaired people could use the company's digital payment platform without seeing the screen. That project benefited more than three million users.

On a different occasion, he was tasked with using computer-vision ML models to identify asbestos fibers. These fibers pose significant health risks when inhaled or ingested and can cause serious respiratory diseases such as lung cancer. Worse, symptoms often don’t appear until years or even decades after exposure.

“It was amazing to use a simple neural net model to save hundreds of hours of human labor,” Ma said. “With that, I saw the potential of ML to influence the whole world.”

To better understand and play a part in that influence, Ma turned to Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program.

Ma was attracted to the MSAI program because of its strong course offerings and faculty who are influential in the industry, he said. He also was impressed by the wide range of roles held by MSAI alumni.

Most notable from Ma’s MSAI experience was Professor Bryan Pardo’s Generative Deep Models course, which covers topics ranging from natural language processing to computer vision. 

“I read more than 20 important papers in this area and discussed them thoroughly in the course with other classmates,” Ma said “It paved a good foundation for me to dive deeper in this field.”

Today, Ma is doing just that as a machine learning engineer at Robinhood, an online brokerage that offers commission-free trading in everything from stocks to cryptocurrency. He secured the job shortly after his MSAI graduation, and he relishes being able to use lessons learned in MSAI to make a big impact for Robinhood customers.

“My specific responsibilities are ML system design, development, and operations,” he said. “I enjoy building a new, useful product that impacts tens of millions of users.”

To stay current with the latest in ML, Ma relies on the skills he learned in MSAI, particularly his time reading research papers on generative deep (neural network) models.

"My work at Robinhood requires research in the cutting-edge developments of ML," he said. "Research courses like Professor Pardo's empowered me to learn fast from the state-of-the-art applications in the field."

Beyond the knowledge learned in the classroom, Ma believes the network he developed in MSAI will be critical for his personal and professional growth.

“I made a bunch of good friends who were helpful and inspiring,” he said. “Such relationships are integral parts of my future career development.”