Theodore Tellides ’23


Tempus Work Environment and Risk Solutions, Concord, NH

Over the last two months I worked at Tempus Risk under Gary Lavoie and John May ’84. Tempus provides advocacy-
based services for injured workers navigating the Workers’ Comp process. I supplied technical skills in order to aid the data analysis process. My work initially involved data cleaning and data visualization using a software called Domo. I built multiple different dashboards each customized for Tempus’s different stakeholders. As the internship progressed, I began to engage in self-created projects. I am a mathematics and physics major and am interested in pursuing a degree in data science. Most of my Williams coursework has been highly theoretical, so I used this opportunity to teach myself the basic applied skills of data science. Such efforts involved learning Pandas and basic machine learning libraries. To my surprise the hardest part of this work was not implementing the machine learning algorithms themselves, but instead building all the pipelines and experimentation processes that had to be constructed around them. My main goal was to try and predict the duration of disability for injured workers. Through trial and error and reading lots of online blog posts, I was able to fine tune the model to use and metrics to plug in.

I spent a lot of time learning to create predictive models using Python.

Because Tempus is only a seed level startup, I was the sole full-time contributor on the tech team. It was my job to interpret my supervisors’ vision and engineer a product that would fulfill their goals. Such a position required me to have both a high level of creativity and a willingness to push back against certain design decisions. These discussions arose primarily about the utility we could squeeze out of our current dataset and the long-term tech structures.

My time at Tempus has increased my enthusiasm for applied mathematics and I am very eager to take Computational Biology in the fall. I also am now planning to take a course in statistics and extend my education of computer science. Tempus affirmed my plan to pursue a graduate degree in data science and I am more interested than ever in the theoretical underpinnings of the field. It was amazing that I could implement a clustering algorithm in only a few lines of code, but I now want to learn the mathematical mechanics fueling the model. Such an interest aligns with my eagerness to help build impactful real-life applications and my love for theoretical coursework. I will continue to work with Tempus throughout the academic year, albeit in a limited capacity. The team was a great source of mentorship, and I am excited to see what they will be able to build in the coming months. Just around the start of school, Tempus will be having an official launch party, and I will be able to meet the team in person for the first time.

I would like to thank Ashesh ’92 and Louise Shah and the 
’68 Center for providing such an amazing and enlightening internship experience. I cannot wait to see what my future with Tempus and in data science will hold.