Becoming a Founder in Data Science/AI x Athletics (and My Step Away from Senior Healthcare)

For the past two months, I have explored the senior healthcare space, focusing on social isolation and its impact on mental and physical health outcomes in the demographic. I was drawn to this space through a personal experience with a family member, and I remain passionate about this problem being solved. However, I’ve decided that this is not the right opportunity for me for two main reasons: 1) there are strong headwinds specific to the problem I was aiming to solve; and 2) I lack experience in the senior healthcare space compared to other industries where I do have expertise.

Why I’m Stepping Away From Senior Healthcare

Regarding industry-specific headwinds, there are two main challenges for building a big business around solving senior loneliness: distribution and monetization. Distribution is notoriously difficult in the senior healthcare space, especially for products where the senior is the end-user. Seniors tend to be less sophisticated in their use of technology and can be less present in traditional marketing channels. Therefore, partnerships with organizations such as senior living communities and community groups are important as a way to get products into the hands of these users. However, the loneliest seniors tend to live at home alone and therefore not in these community groups. The gap in technical fluency also presents an additional challenge: the product must appeal to two different audiences, specifically the senior and the person assisting with setup, such as their adult child or a professional caregiver. At the same time, monetization for this type of product can also be difficult, since loneliness is not a medical diagnosis. Loneliness and mental health can be relatively difficult to measure compared to other medical ailments (e.g. blood pressure or cholesterol), so willingness to pay for unmeasured impact can be lower. Interestingly, non-profits do a lot of amazing work in this space; however, entering a space currently led by non-profits and attempting to create a revenue-generating business isn’t promising.

Second, my lack of experience in senior healthcare was also a challenge. I did not have a refined intuition for how to navigate this space, which caused me to slow down at several points in my exploration. I am certain that I could have built a deeper understanding of senior healthcare over time, but, given time constraints, this was not an investment I wanted to make at this point in my startup journey.

My time exploring senior healthcare was not wasted; I was able to explore a new industry with a completely fresh perspective and dig into the dynamics of the industry without any preconceptions. At the same time, I learned that I want to work in a field where I have both passion and expertise. This leads to the next stage in my founder journey — building in the intersection of Data Science/AI and Athletics.

Looking Forward: Data Science/AI and Athletics

I’m excited to build in these two areas because I have truly felt the pain of the users that I’m building for. My combination of experiences gives me the perfect Founder-Market Fit for building here.

I have worked as a Data Scientist at a premier tech company (Twitter) and at two early-stage startups. I understand how to apply the principles of data science in both a scrappy startup environment and at scale in a world-class data science organization. Perhaps more foundationally, I double-majored in Economics and Psychology at Yale precisely because it was the best combination of majors to help me understand how humans rationally (and irrationally) make decisions. Since AI is fundamentally just leveraging existing datasets to detect patterns in novel data, my Data Science background also gives me an edge in deploying AI to solve complex data problems. In addition to my data science background, my experience in engineering allows me to test potential solutions in the real world, instead of just exploring them conceptually. Building an end-to-end prototype can help to investigate hypotheses and determine the actual utility of a product, reveal a customer’s willingness to pay, and indicate the specific use cases that matter most to a customer.

I have also been building an expertise in athletics since I was very young. I started playing ice hockey at age 3 and lacrosse at age 8. I went on to play Division 1 lacrosse at Yale, where we won a National Championship and I was twice named an All-American. I also played professionally for two years post-college in Dallas while working in my first full-time data science job in New York City. I hung up my cleats in 2019 but have remained obsessed with athletic improvement today, mainly through triathlons and running races. My next race is the San Francisco Half Marathon next month, where I’m aiming to beat my previous half marathon PR of a 6:45/mile pace.

What’s Next?

I’m excited to lean into these areas of expertise that I have built up over the course of decades. I’m currently exploring and building around a few ideas:

1. Providing measurement and detecting proper mechanics for lacrosse “wallball” — passing a lacrosse ball repeatedly against a wall. This exercise improves a player’s ability to throw and catch a ball with precision. A solution here would solve a long-term personal pain point, as I played wallball almost every day throughout my career but never had a great sense of whether my wallball metrics were improving. Using computer vision, a user could track metrics on each wallball session (number of repetitions, successful and failed passes, breakdown of types of passes, etc) while also receiving real-time feedback on their form. The same concept could also be applied to a player shooting the ball toward a net, with the ability to measure accuracy, detect tendencies, and approximate shot speed.
2. Finding value in niche sports betting markets and creating market intelligence using data science techniques. Many college/pro teams have more advanced data that has been tracked during past seasons that is not particularly useful to their program today. However, this data could feed into a model to understand statistical trends that contribute to winning/scoring more goals, etc. Alternative data sources could also be leveraged: satellite data could be used to detect daily activity on team practice fields, and school calendars could be used to match up finals period or school breaks with game schedules.
3. Generating real-time feedback on exercise movements, such as lifting weights or rehabbing from an injury. In discussions with physical therapists, I’ve found that this could be particularly useful for providing guidance for patients doing at-home exercises, away from the watchful eye of the PT. Similarly, during weightlifting sessions, I personally welcomed feedback on my form, particularly as the weight got heavier. However, I would sometimes lose track of the number of repetitions because I was entirely focused on my form; using computer vision to track exercise details such as the exercise, weight, # reps, rest period, etc. would reduce the cognitive load of a workout.

Interested? I’d Love to Chat!

I want to collaborate with others on these projects, and I ultimately want to find a co-founder to build a company with in this space. If these projects excite you, or you know someone who would be excited by them, please send them my way. In the meantime, I’ll be jamming on ideas, talking to potential customers, and building prototypes!