Senior Data Scientist — Kyle Florence

This week’s spotlight is on Kyle Florence, PT, DPT, SCS, a non-clinical physical therapist who is now Senior Data Scientist for Blue Cross Blue Shield of Massachusetts!

What is your full name, title, and company name for your current, primary role?

Kyle Florence, PT, DPT, SCS – Senior Data Scientist at Blue Cross Blue Shield of Massachusetts

Where are you located?

Massachusetts.

Where did you go to PT school, and what year did you graduate?

University of Massachusetts Lowell, 2018.

What did you do when you first finished school, and for how long?

I initially worked for Professional PT as an outpatient PT for about a year. Then, I moved down to Connecticut and worked at Connecticut Children’s as a sports PT for two and a half years.

In what setting(s) did you work, and what types of patients did you treat?

I worked in outpatient and specialized in sports. Being at a children’s hospital meant my main population was athletes from middle school through college.

What did you enjoy about your early roles? What didn’t you enjoy?

My time at Professional was like drinking from a fire hose—but I think it was quite valuable to see tons of patients with all sorts of presentations early in my career. Along with the patient volume, Professional offered the Institute of Manual Therapy (IOMT), where I learned hands-on manual therapy techniques and clinical reasoning from Martin Langaas. I firmly believe the vast amount of hands-on education, paired with a large patient volume on which to practice early in my career, was key to becoming an effective clinician.

After leaving Professional, I found myself on a new team of evidence-based clinicians at Connecticut Children’s. My colleagues there were a mixture of board certified clinical specialists and clinicians performing peer reviewed research on cutting edge treatments. The team here was a great motivator to study hard and become board certified, as well as constantly improve my care through staying current with the latest papers.

This love for scientific literature ended up being a strong motivator for my current path in data.

I can’t think of much negative, other than the typical outpatient PT stuff. I really didn’t like overlapping patients, which ends up being the norm in outpatient.

What else have you done since then, prior to your current role?

My first opportunities outside of the clinic were a data analyst role at Pratt & Whitney, followed by a data scientist role at Alight. 

When and why did you decide to do something non-clinical?

During the pandemic, my hours were cut, and all of my friends were working from home. I decided to look into insurance or utilization review type roles—but so was everyone else. Every role had something like 200+ applications within hours of posting on LinkedIn. I decided to drop that pursuit once things opened back up in the clinic.

Initially, things were going very well. We were treating one patient at a time due to spatial restrictions. I was studying for my board certification, so my care had never been better.

Soon thereafter, we were back to dovetailing patients, and I started to burn out hard.

I decided to look back into non-clinical options around this time. I expanded my search to analyst positions, as I had also been sharpening my programming skills on the side.

This was when I finally started getting interviews and jumped at one even though it had nothing to do with medicine. This was actually a benefit to me at the time.

What are you doing these days?

I’m a data scientist at BCBSMA! I was brought on to work on clinical modeling because of my technical, as well as clinical, background. At the time of writing this, I am brand new at this company, so I don’t have a lot of details, but I am excited to improve the healthcare experience from the payer side.

Getting the right person to the right care as soon as possible, ideally before things are problematic (predictive modeling!), is such a fun challenge, and I’m fortunate enough to get to work on these problems.

Are you still treating patients, or are you solely non-clinical?

I don’t treat any patients, other than my own body that is constantly falling apart.

How long have you been in your current senior data scientist role?

I started in April 2024.

What do you wish you would’ve known before going into this senior data scientist role?

The data in the real world will always be much messier than the data you learn on!

Did you get any special certifications or training along the way to help you get into your current role?

So, this one is interesting. I did get my master’s in Computer and Information Technology from Penn. However, I got my first position as an analyst before I was admitted to the master’s program. When I made the jump to my first data scientist role, they really only cared about my experience.

I do think a degree is becoming more and more necessary, as the job market gets more competitive. Hiring managers can simply use it as a filter and still get hundreds of applicants.

There are cheap ways to get an online bachelor’s degree these days, as well as a master’s degree, if you have more experience. Once you have your foot in the door with your first role, it may become less necessary. However, without a foot in the door through incredible networking or something like that, I recommend a degree.

How did you find your job? Did you apply or find it through a connection?

For my first data scientist role, I posted in a group offering to help someone, and someone else messaged me later asking if I’d be interested in hearing about a role. It was a “right place at the right time” moment, and I am very thankful 🙂

My current role was a simple application and interview process without networking. 

What was the interview like for the senior data scientist role?

Interviews in the data science space almost always include some sort of technical assessment.

This was new—it would be like going to a PT interview and having someone ask you to write out a treatment plan for someone three weeks out from an ACL reconstruction (as they probably should).

In data science, they ask you to code on the spot or give you take home assessments which tend to be fair but stressful. Along with technical assessments, you might get questions like how to set up an A/B testing experiment, what metrics you would measure for certain business problems, etc.

How have people reacted to you leaving patient care?

I remember really feeling like a failure. I had worked so hard to be an excellent clinician and just jumped into a profession I frankly googled how to do. It seemed like such a crazy move, and I hated talking about it. But! Most people actually seemed to say they saw it coming. I had been programming and automating stuff on my personal computer since high school, so I think people were confused why I wasn’t in this space already.

What’s a typical day or week in the life like for you? What types of tasks and responsibilities fill your time?

Every day is a bit different, depending on the project I’m working on.

Most days start with some kind of stand-up meeting, where the team meets and discusses anything blocking their work. Some other meetings might be on the schedule, where I’m talking to stakeholders about model performance, current data needs, or new model ideas.

Between all of this, I’m playing around with the data! I generally go looking for the right data by querying different databases with SQL, then pick it apart and start to model in Python. The remaining time (and occasionally full days) is spent debugging issues with older models and deployments.

What are some of the rewards of your role? What are the biggest challenges?

The best reward is seeing a customer or end user impacted by my work in a positive way. This might be getting the right message to the right person at the right time to save them some money or improve their healthcare.

The biggest challenge is working at a large company and some of the clutter that comes with that. There are times where I create a model, and it needs to go through five other teams before it reaches a customer, so it can be hard to see my own impact in it all.

How did your clinical background prepare you for this role? Which skills transferred?

Lots of the data that I am able to use is medical history and claims data. Having my clinical background has made me a quick subject matter expert on these use cases. I am able to pair typical data-driven approaches with my clinical knowledge to speed things up and improve the performance of the models.

Along with this technical stuff, being a clinician helped me be comfortable in the gray spaces of my work. Much like in medicine, where not every treatment will turn out the same way, the same applies when you run a model on a whole population. You’ll get false positives and false negatives, and you need to be able to investigate and improve where you can.

What type of person do you think would do well in your senior data scientist role?

Data scientists need to be technically sound in statistics and able to perform the analysis and modeling via code.

The person who will do well is a lifelong learner, as the technical landscape is always changing, and relentlessly curious.

There are times where you may not have a clear task, and you are given some time to explore. Curiosity will drive a good data scientist to constantly try to find new things and improve what’s out there.

Do you work remotely or onsite?

I work remotely four days per week and go into the office one day per week.

Did you read any books, take any courses, or do anything special overall to get you where you are today?

In order to get good at data analytics, you need to practice. You should try to create projects that help with your daily life, like analyze your credit card spending, automate a workout tracking spreadsheet, etc. These types of things were how I started, and I would constantly look for YouTube videos to get them going and improve them.

Once I had some basic data analytics knowledge, I read this free book that really drives home the basics of data science and statistical learning.

What is a typical career path for someone in your senior data scientist role?

I think most data scientists start out as data analysts or business analysts (honestly, any analyst title can overlap a ton). Here, you’ll get a good sense for using data to drive change and solve business problems. From data scientist, you can continue as an individual contributor or move into management of some kind.

What is next for you? What are your high-level career aspirations?

I’m loving this data space right now. I could see myself moving around between data engineer, machine learning engineer, or stay as a data scientist.

My expertise is definitely in healthcare data, so I think that’s where I’ll be able to have the most impact. One day, I’d like to lead a team of scientists to create models that help people get better healthcare, or improve clinicians’ abilities to provide better healthcare.

Ready to launch your own non-clinical career?

What would you recommend to someone who is considering going into a role like yours? Do you have any special words of wisdom for the readers?

Look—data scientist is a technical role. Unless you have a math undergrad, you’ll need to learn a whole new skill set that was not taught in school. It can be super intimidating, so start slow and learn the basics of analytics first.

Start by getting really good with Excel, then SQL, then Python. From here, you can start learning about the statistical methods that will serve as the foundation for your machine learning and modeling future.

Once you have the technical skills, you need to market yourself well. An important distinction you’ll need to make is that you are a data scientist who happens to have a rehab background, not that you are a rehab professional who happens to be a data scientist. Employers want data scientists with subject matter expertise. This will require some creative writing on your resume 🙂

What would you teach to today’s graduate students in your profession, if you had the opportunity?

We did learn about this in school; it didn’t click as to how important it would be at the time, but, STATISTICS! Even if you aren’t going to be a data scientist, we as rehab professionals need to be SO much better at understanding the statistics that drive our profession’s research papers!