University of Toronto alumnus Shiva Amiri, formerly senior program lead at the Ontario Brain Institute, is director of data science at Zymergen, of Emeryville, Calif. Zymergen has been recognized as a leading company for female hires in Silicon Valley.
Of the more than 400 people at Zymergen, about 40 per cent are women, a relatively high number for a tech company. I’m the most senior woman in technology.
I work with teams with a data science vision – what we need to focus on and get the most from data we produce. Our clients – Fortune 500 – care about products and materials. We build genetic strains to build materials using machine learning and data science.
There are many applications and different kinds of data coming together that enhance products, pharmaceuticals, drugs, farming, agriculture, paint … anything. The data science team I head has 11 people, six women – that’s rare.
I haven’t felt a “bro” culture here. Have I felt that before? Definitely. My attitude was to ignore the noise, head down, work to overcome obstacles; if you’re good, it will make a difference and get noticed. There are times to say “it’s not making an impact or these guys aren’t good, this team isn’t good” – move on.
I was born in Tehran, in Iran until I was about 10 years old, moving to Canada in 1990. I wanted to do something useful for humanity – as cheesy as that sounds. I started in human biology to be a doctor. Realizing genetics was picking up, data was big and computers would play a role in medicine, I did a double major; University of Toronto was the only university to let me.
I did an undergraduate essay on why women don’t go into computer sciences much – women like to see their work applied. “Can this be applied to linguistics or can I use this to enhance disease diagnosis or help the elderly?” It’s cultural. In Russia, a lot more women go into engineering, the numbers are also much better in China and Iran.
In summers, I worked in a bioinformatics lab then went to Oxford University for a masters, switching to a computational biochemistry PhD. I focused on large-scale computing with bio-related data in neurological neuro-muscular diseases using computational methods to solve problems. I finished my PhD in 2007, co-authoring Oxformed: A Journey Through Oxford, about being international graduate students. I wanted to go back to Toronto, but there weren’t a lot of tech companies or start-ups then.
The first job that made sense was program officer at the British High Commission; I was there four years. It gave me different perspectives on communication and scientific diplomacy, but I wanted to get back into actual science.
I joined the Ontario Brain Institute to look after some neuro-technology programs and got on the Brain Code project, developing a large-scale platform for the storage, management and analysis of images, genetics data and clinical data across hospitals. There were many challenges; collecting data into a platform, privacy and security, governance, getting people to understand why collecting data was important. Hospitals sit on very valuable data that could enhance discoveries, push research. We funded research, asking hospitals to share it with colleagues to relate it to diseases they worked on. I became Brain Code’s manager, then [was] ready for something else.
I’d worked with start-ups but was missing start-up experience. I joined New York’s Real-Time Data Solutions as chief product officer in genetics data. I thought we needed to go beyond and analyze multiple kinds of data; we spun off BioSymetrics and I became CEO. We developed two great products that won awards; one, an integrated machine learning on biomedical data, [the other] a finance product.
At first, there were many skeptics. We got very good results based on predictions we made working with a hedge-fund partner. Our models were making a lot more money than the fund. As we kept building and the numbers kept coming in, we said, “This has to stop. This can’t be happening.” We’d done things simply, a robust job of tackling a targeted problem.
To me, artificial intelligence means when computers imitate human behaviour. Machine learning is a subset, one way to do artificial intelligence is based on lots of data. If we can train statistical models then we can make predictions, as a human would. It’s making an impact in multiple sectors; retail, automotive, self-driving cars, farming, medicine and biology. For me, it was a no-brainer.
We’ve launched a new weekly Careers newsletter. Sign up today.