Aya Soffer: IBM’s AI trailblazer

Soffer has made an impact as IBM’s vice president of AI - let’s take a look at why she made our list of Israeli gateways to the space of Artificial Intelligence

CTech 12:3406.01.21

TAGS:

Name: Aya Soffer

Title: Vice president of AI at IBM

Notable facts:

Worked for two years at NASA
Main focus is on IBM’s Watson AI tools

Aya Soffer, Vice president of AI at IBM. Photo: Miri Davidovich

Twenty-one years ago, Aya Soffer joined IBM and started framing the ways we would one day interact with the technology around us. As Director, Information Management Analytics, Soffer would work up the ranks and one day become the VP of AI for the entire company - but how did she get there?

Soffer attended the Hebrew Reali School before obtaining a Bachelor of Science from The Hebrew University in Jerusalem. These formative years in the 1980s would set the foundation of the technology ahead of her and provide the opportunity to venture across the Atlantic Ocean. At the University of Maryland, she earned her Masters of Science and Ph.D., each in Computer Science.

For two years after her education in the U.S. Soffer worked as a Research Associate at NASA, specifically the Goddard space flight center where she worked on digital libraries for earth science data. It was these years, in the early 1990s, where she studied the ‘Retrieval by Content in Symbolic-Image Databases’ for her thesis. With an education and early career spanning decades and continents, she was ready to join IBM.

“These days my main focus is on Watson,” Soffer explains. Watson is IBM’s question-answering computer system that can answer questions posed in natural language. “Where will Watson go next? Augmented Reality, Affective Computing, Debating technologies and more?” she teases.

Of course, there is more to IBM than just Watson. As leaders in the field, Soffer is spear-heading the company in helping understand unstructured data, including image, video, speech, text, and multi-modal data. It may end with the consumer enjoying IBM’s AI tools, but it first means IBM teams work on the ‘entire spectrum of the problem domain’ - this includes market analysis, product definition, and solving the toughest technical challenges. According to Soffer, she also “works on shaping IBM Research strategy in this area and with the corporate in general on future directions for the products.”

How does IBM intend to utilize AI? Soffer is working on a myriad of AI solutions, such as customer service, financial tools, cybersecurity, healthcare, or video. With each decision comes the fundamental properties for ethical AI at IBM: explainability, fairness, robustness, transparency, and privacy.

In 2019, Soffer spoke with CTech at the Mind The Tech TLV conference where she discussed how the team ‘moves the needle’ on what is coming next with AI. “one major area is the mastering of language - it’s an area that still has some way to go… we’re also looking at what it takes to trust AI.”