DSI Graduate Uses Precision Medicine to Improve Healthcare in Argentina

Diego Llarrull
Diego Llarrull

Every day as part of his job, Diego Llarrull uses data science for good. The work he does allows doctors to diagnose children with rare diseases and helps women with breast cancer to receive treatment.  Llarrull is a Data Science Analyst for Héritas, a precision-medicine company in Rosario, Argentina.

He was born in Argentina, where he studied computer science at the Universidad Nacional de Roario. While there, he was awarded the Argentine Presidential Fellowship in Science & Technology, which allowed him to study for his master’s at the Data Science Institute. As as part of the fellowship, he agreed to return to Argentina after he earned his master’s. Luckily for him, just as he finished his degree, Heritas was looking for a data scientist. It was a perfect fit, and he got the job.

In this Q&A, Llarrull talks about working for Heritas, his studies at DSI, as well as how he uses data science for societal good.

*****************

Can you talk about how you returned to Argentina for your job?

As mentioned, as part of my fellowship agreement I had agreed by contract to return to Argentina once I had finished the master’s in data science. Consequently, and right before sitting for my final exam, it came to my attention that Héritas, a precision-medicine company from my hometown was looking for a data scientist, one who could work on machine-learning models for the prediction of pathogenicity in genetic variants (that is, mutations in our DNA), as well as techniques to visually explore the vast cohort of genetic variants from already sequenced patients.  

You returned to Argentina quickly.  Did it all work out well?

I returned to Argentina in late September 2016 and by mid-October I was already working at Héritas. It’s a thriving startup born from two very reputed companies in Argentina. Within this scenario, the amount of workload and the ability to be constantly switching contexts from work to study that I developed while at the DSI program truly paid off for me.  

Can you talk a bit about your work?

As a Data Science Analyst, I help develop machine-learning models for mutation classification and Natural Language Processing tools that help to locate relevant scientific information for easing diagnosis.

Do you use your DSI learnings in your work?  

Yes. All the time.On any given day at my job I find myself tweaking parameters of a family of machine learning models, helping oncologists visualize disease patterns in our local population, designing custom apps to improve data-generation within the company, or even designing workflows to maximize the company's throughput.

Did you find the master’s program at DSI to be challenging?

I recall the exact moment when Jonathan Stark, then the director of operations for DSI, while leading the Pre-Orientation online session for the MS in Data Science, firmly stated: this program is “HARD.” Having graduated from a very reputable undergraduate program (equivalent to an MS in Computer Science) in Argentina,

I thought I knew what he meant. It turned out I didn’t: the MS in Data Science proved to live up to Columbia University's reputation as a top-level institution.  

The intense workload, together with the enormous amount of available seminars and guest talks, as well as two-day long Hackathons, demanded a great level of effort from me. Thankfully, this was leveraged by the great support of all professors and, most of all, my classmates, who quickly set me into a pace that I could have never thought I had in me.

What are your best memories of DSI?

From my time at Columbia, I recall excellent professors like Eleni Drinea, and very demanding courses like Statistical Inference and Modelling and Machine Learning for Data Science. I most recall the warm kindness, constant support and astonishing intellect of not only the DSI staff, but also of my classmates, with whom I spent endless days and nights solving statistical problems on a sheet of paper or optimizing machine-learning models to yield an infinitesimal improvement over our previous effort.

Which of your achievements are you most proud of?

In March 2016, a team of DSI students won the Microsoft Award at DevFest 2016 for creating an app that showed the emotions of presidential candidates' debates on TV(https://devpost.com/software/debate-in-emotion). It was a side project we worked on. Also, for my Capstone project, I was part of a team that worked on a medical imaging project in which we used deep learning techniques. That project also turned out very well.

Is your work having a socially beneficial impact on Argentina?

The work is helping many patients in Argentina; from children with rare diseases in need of accurate diagnosis to women with breast cancer seeking genetic explanations for better treatments. Our latest service seeks to prevent invasive testing of women during the initial weeks of pregnancy; and our non-invasive method is more accurate in discovering trisomies like Down Syndrome or Patau Syndrome or even in establishing the sex of the baby than traditional (ecography) methods. We are confident that as time passes and awareness of the benefits of these services becomes more widespread, our services will reach people irrespective of their economic situation.


--By Haniya Javed



550 W. 120th St., Northwest Corner 1401, New York, NY 10027    212-854-5660
©2018 Columbia University