Hoda Eydgahi always thought she’d be a physician — so much so that she got her undergraduate degree in biomedical engineering. However, despite graduating a year early, and as a valedictorian, from Virginia Commonwealth University, she was initially rejected by medical schools. So she decided to apply to MIT’s Department of Electrical Engineering and Computer Science with weeks to go before the deadline — and got in.
It would prove to be the start of a “random walk” over the next twelve years, bouncing back and forth from the theoretical to the applied, from raw startup entrepreneurship to scouting out new startups, and to where she is today — a data science manager at technology-driven fashion retailer Stitch Fix and an angel investor in early-stage startups.
On arriving at MIT, Hoda first completed a Master’s degree with MIT Media Lab Professor Rosalind Picard, developing wearable sensors and analytics for medical and health uses, such as monitoring stress and seizures. Still fascinated by biology and health technology, she considered medical school again. Despite being accepted that year, Hoda decided to stay at MIT instead. “I went to a few medical school interviews. I was sitting in the waiting room and thinking about how everyone’s constantly competing with each other from day one. First it’s for a seat at the school, then after that it’s for residencies and fellowships — it’s competition, competition, competition.”
Hoda contrasted that with her first years at MIT, where she said she was “clobbered” but well-supported: “I was surrounded by incredibly smart people who are much more brilliant than I am, but everyone’s so humble. There’s no ego. All the grad students help one another out, the professors are helpful — it makes for a very healthy environment.” What’s more, she felt she wanted a solid mathematical foundation for whatever came next. So, she went back to the classroom.
During her doctoral studies, Hoda was co-advised by LIDS Professor John Tsitsiklis and systems biologist Professor Peter Sorger of Harvard Medical School. She used Bayesian methods to study apoptosis, or programmed cell death. Hoda explains the significance of her work: malignant cancer cells don’t die when they should, and when their growth spirals out of control, that’s when cancers form. Understanding the normal cell death process can shed light on potential targets to treat cancer and help predict how cells will respond to different drugs. But many biological processes are hard to observe with current technology. “We used probabilistic models to determine which proteins interacted with which other proteins. If you think of it as a graph, each protein is a node. But the state of technology isn’t sophisticated enough to tell us whether protein A connects to protein B or protein C, so we used probability to infer the shape of that graph.”
While Hoda was doing her PhD, the “random walk” took yet another direction. Like many PhD candidates, Hoda had entrepreneurial aspirations and wanted to have an impact beyond her thesis. To explore this interest, she took a class in entrepreneurship at MIT’s Sloan School of Management and entered the annual MIT 100K Elevator Pitch competition with three of her classmates. Here they pitched the social enterprise Bluelight, which enabled lowincome households in the Middle East to access microfinance savings accounts via their mobile phones — and won third place in the competition. The idea also won the grand prize in social enterprise at Harvard Business School’s New Ventures competition that spring. Following the momentum of this success, Hoda spent her first year after graduation helping her three partners build Bluelight in Jordan, where one of the founders was from. But burnt out from the demands of building a startup as a first-time founder, she decamped to the San Francisco Bay Area.
In San Francisco she was drawn to another, now larger startup, Color Genomics. The company produces affordable genetic tests for ovarian and breast cancer, and brought Hoda back to her combined interests in medicine and data science. At the time, Color Genomics was a ten-person startup and Hoda joined as their first data scientist. She wore many hats and did production engineering on the fly, shepherded algorithms, and wrote software to facilitate the product’s launch in 2015.
All of these experiences — a strong foundation in probability and data science at LIDS, product launches at startups, and production software engineering — fed into Hoda’s next move as a data scientist for Stitch Fix.
Hoda was one of the earlier members of a team now consisting of over 100 data scientists. She explains how the company uses both personal stylists and algorithms to send customers ap- propriate picks, based on their preferences, geography, feedback and a host of other factors. It even uses algorithms to design clothing.
“Data science is really stitched into the fabric of Stitch Fix,” she says. “From the very beginning, we had a chief algorithms officer. That enables the data scientists to have a lot of scope and responsibility. We’re not just focusing on building algorithms, but also responsible for productionizing it.” Another thing she likes? “Diversity. [In data science] I’ve always been used to being one of the very few women in the room. But at Stitch Fix, that’s not the case.” More than half the company is female and it was the only tech IPO led by a female CEO in 2017.
The LIDS network, too, has been a “gift that keeps on giving,” she says. Many LIDS alumni end up in academia and research, and she’s been able to work with her LIDS colleagues’ graduate students to fill data science positions and other roles.
Today, Hoda seems to be at a happy medium. Beyond her day job, she is a scout for venture capital firm Sequoia Capital, looking out for exciting early-stage startups. She also enjoys hiking and cycling, travel, and photography in her spare time. Perhaps the “random walk” of Hoda’s career is not so random after all — as clearly, it’s trending toward success.