The Crunchbase “Female Founder Series” is comprised of stories, Q&As and thought-leadership pieces from glass-ceiling-smashers who overcame the odds and are now leading successful companies, or investing in them
This article is part of the Crunchbase Community Contributor Series. The author is an expert in their field and a Crunchbase user. We are honored to feature and promote their contribution on the Crunchbase blog.
Please note that the author is not employed by Crunchbase and the opinions expressed in this article do not necessarily reflect official views or opinions of Crunchbase, Inc.
Grace Chang, a Berkeley-based entrepreneur, has spent the past three decades quietly building frontier technology across many of today’s preeminent security, enterprise, consumer, health and fintech companies. In this article, Chang shares her story and how her company, Kintsugi, a voice biomarkers software startup, is tackling access to mental health care head-on.
I started out life on difficult footing, with my family emigrating from Taiwan to the United States with modest means. This is an experience I share with my future co-founder, Rima Seiilova-Olson, who grew up in a yurt without running water or power in Kazakhstan. Both our families struggled to put food on the table, yet we managed to prioritize education above hunger, strife, alienation and even hope.
Everything related to mental health was taboo in our cultures. Today, although we have terms like boundaries, conditional love and along the spectrum to describe our childhood experiences, at the time, I couldn’t have felt more different from my peers. When I was invited to the community pool in first grade, I put together a neatly stacked set of hand drawings and sold them door to door until I reached the 75 cents needed to pay for the day’s entry.
That difference grew when, in 2000, I attended University of Southern California on a scholarship to further my deep interest in computer science. Of my 230 classmates, only two were women. Often, my late Friday nights working with lab partners on assignments were anxiety-ridden with imposter syndrome, body dysmorphia and a fear of unwanted sexual advances. Few of my male counterparts wanted to invest time working with the “minority card” candidate who, as some put it, “likely only got in because of her gender.”
Worse still, when joining founding teams in Silicon Valley, select executives would often devalue my work and tell me “stay in your lane and act more subservient in order to move up and for us to really like you.” Relentless work hours and fitting in while providing an opinionated stance without being seen as the token woman or POC, was a continuous struggle that felt more personal as I advanced in my career—and ultimately led to burnout.
The culmination of events reactivated painful childhood memories that I hadn’t yet dealt with. However, when seeking mental health support through my provider, I struggled for nearly half a year to access any professional help. Upon referral from a close friend who was similarly working in tech and a POC, I was able to connect with an extraordinary therapist to unpack my experiences and navigate the present. They understood what it meant to feel like an outsider, as well as the unique dynamics of an impoverished immigrant family. At the time, I was so desperate for help that the $450 per hour out-of-pocket rate became a necessity that I had to budget for, a luxury I knew many did not have.
The dejected sense of self was a pain I never wanted others to feel. When I met Rima at the OpenAI Hackathon in San Francisco, we bonded over similar challenges we had accessing mental health care. There was nothing I wanted more than to dedicate my imagination and ambition to help people going through a similar state: feeling burnt out, taken advantage of and disposable.
Rima and I came together to embark on a new startup of great personal significance and to build a unique product to serve this massive gap: We would use our experience to build sophisticated technologies in other highly regulated spaces and pour it into an area that was much more personally meaningful. With our AI and machine-learning expertise, we built a bridge for everyone to access mental health care, starting with faster detection by the organizations they trusted.
Searching for a voice in mental health
I kept a journal throughout middle and high school, and understood why therapists recommend it as part of patients’ treatment. Pennebaker’s paradigm from the 1980s showed how articulating top-of-mind issues help people mitigate stress, anxiety and depression. Now, with the pervasiveness of Alexa, Google Home and smartphones, I hypothesized that people could turn to voice-based journaling to vent and connect with others. So, I launched the first voice journaling app in the Apple App store shortly before incorporating Kintsugi. But despite being a valuable tool, that was only step one.
We have Apple watches and Fitbits to help us track our physical fitness with steps; however, there isn’t anything analogous for mental health. I’ve spent my career unearthing data for unique insights into consumer behavior, so it was obvious to me that you can’t improve something that you don’t measure. Yet, to this day, we measure mental health with antiquated paper Patient Health Questionnaire-9 (PHQ-9) and General Anxiety Disorder-7 (GAD-7) tests.
In the fall of 2018, MIT produced a paper on what was then the state-of-the-art machine-learning model to detect clinical depression from short clips of speech. At the same time, our team created the world’s largest voice dataset-specific for machine learning and mental health, thanks to the hundreds of thousands of people across the world using our journaling app. This gave us the luxury of being able to carry out thousands of experiments on clinical depression and anxiety detection in real-time.
We discovered that it wasn’t what patients said that determined clinical depression or anxiety, but rather, how they said it, regardless of language, gender, socioeconomic or ethnic background. This was a startling find.
Listening to people’s needs and bridging the gap
People act, react, look and sound different when they’re depressed or anxious. Understanding those signals can go a long way in destigmatizing mental health, and in getting patients solutions for their well-being at the time of need. The first step in treatment for mental health isn’t a psychiatrist, but to provide information on a wide array of solutions that can prevent a condition from escalating; tools from digital health and promoting well-being to telecoaches, therapists, nutritionists and sleep experts. Patients must feel supported in their care journey in both mind and body.
As we brought Kintsugi to life, I interviewed clinicians, nurses, therapists and psychiatrists; I saw in them an incredible passion for their patients, despite an often strained system that hasn’t changed in decades. These same interviews and deep research and development yielded our team’s initial funding from the National Science Foundation in 2019, and multiple prestigious Small Business Innovation Research (SBIR) grants for novel AI technology. All of which helped accelerate our efforts in tackling the impending tsunami of the mental health crisis triggered during the COVID-19 pandemic.
We wanted to alleviate the jam at the front door of health care by using our novel technology to provide mental health visibility to clinicians so patients are treated holistically and in a timely manner. At a time when people’s health care was largely going virtual, we decided to create an intelligent layer on top of clinical calls coming in and out of call centers and telehealth platforms. That layer could detect depression or anxiety in patients’ voices for clinical providers, improve screening for all members and, in doing so, connect patients with support at their time of need.
Looking ahead
To build our machine-learning models, it was critical for Rima and me to fold in our diverse upbringings, cultural frameworks and life experiences to create products for scale. Fundamentally, we believe diversity of perspectives improves the usability of products, and as women and persons of color, our life journeys made us acutely aware of the hardships people of other cultures face when accessing health care. It has also made us more sensitive to mitigating bias in our machine-learning works and redoubling efforts to recruit candidates with diverse experiences.
Technology has the power to streamline processes that save lives every single day. AI tools like Kintsugi have the potential to reach hundreds of millions of people over the next decade, removing the burden on our dedicated doctors and nurses so they can concentrate on providing quality care.
Everyone in the health care space has a responsibility to address the mental health crisis, be it by integrating support tools into their offerings or by continuing to uncover new ways to shine a light on these less-visible conditions and connecting patients to solutions. Kintsugi, the art of repair, has been a guiding principle for our organization. We strive to be proponents for those who are in need of human support. No one is ever too broken to be loved and cared for, and in the process of restoration we hope to make our collective lives more beautiful.