I am a Ph.D. Candidate at MIT Sloan in the Information Technologies (IT) group - a subgroup of the Management Science program. I am also a research affiliate at the Stanford University Digital Economy Lab (S-DEL) and the MIT Initiative on the Digital Economy (IDE).

My research focuses on the Future of Work and how information and automation technologies transform society. My main advisers are Erik Brynjolfsson and Sinan Aral. I am applying to academic positions during the 2021-2022 job market cycle.

In my job market paper, I study how information and automation technologies have changed (and will continue to change) the skill compositions of occupations and what those changes imply for the values of skills and other human capital.

Leveraging over 200 million US job postings that were posted on online job platforms over the last decade, I find that low and medium-wage occupations’ skill compositions changed more than those of high-wage ones, implying a ‘double whammy’ for workers in those types of jobs: not only are they at higher risk of displacement through automation, they also need to reskill and learn more new skills in order to stay productive and keep their jobs. In this work I also highlight the importance of using compositional data analysis (CoDA) when working with panel data from platforms or other data in which the number of observations grows significantly over time. Finally, I derive a panel of market, i.e. skill demand-based, returns to skills, which should be useful for other researchers, job seekers, and employers alike.

Relatedly, I’ve studied how exogenous shocks such as the unprecedented Covid-19 pandemic (Brynjolfsson, Chi, Jin, Steffen, Bai (2021)), and digital data breaches (Bana, Brynjolfsson, Jin, Steffen, Wang (2021)) affected firms and their labor, skill, and IT demands. In both projects, a deeper understanding of firms’ specific occupational and skill demands through their job postings contributed to a better understanding of firms’ response and resilience to shocks.

I’ve also used modern Machine Learning and NLP tools on the raw text of job postings to study the language of jobs to help job providers (and seekers) write (and apply to) better and more suitable job postings and thereby improve the labor market matching efficiency. Specifically, for our working paper ‘Job2Vec: Learning a Representation of Jobs’ (an early version is available upon request), Sarah Bana, Erik Brynjolfsson, Daniel Rock, and I use contextual embeddings from BERT to encode job postings and predict their occupational labels. While our test accuracy of over 60% is impressive (considering it’s a multilabel classification problem with nearly 1000 distinct occupational labels), the main scholarly insights come from (i) the models’ softmax layer, as we can use its probability distributions to identify job (dis)similarity and specialization and (ii) text injection experiments in which we alter the job posting texts in predefined manners, such as translocating job postings to other cities or additing additional skill demands or remote-work opportunities.

In 2018, I was also a Data Science Intern in LinkedIn’s Economic Graph team in San Francisco. In ongoing work with them I study how workers’ career paths and transitions have changed over time. In a future project I plan on using M&A’s to further study career paths, network connections, and homophily between workers of the acquiring and target firms.

Before joining MIT in 2016 I was an RA for two years for Susan Athey and Markus Mobius at the Microsoft Research (MSR) New England Lab and helped on the Google News and Indonesia Immunization papers among other projects.

Besides research, I also enjoy teaching. I have received the MIT Kaufman Teaching Certificate and have been a TA for the MIT Sloan Analytics (A-)Lab for several years. I believe that the class’ concept of matching students with firms to help these firms use their data to answer real business poblems is a win-win, and I’ve found my work in matching and mentoring students as well as meeting with companies to refine their project proposals extremely rewarding. I have also TA’d Erik’s Ph.D. Seminar on the Economics of Information Technologies, as well as several Executive MBA classes, including the Global Organizations (GO-)Lab. Finally, for the past 3 years, I have coorganized Erik’s Digital Economy Lunch Seminars at the MIT IDE and Stanford S-DEL.

Finally, I’m an avid runner and pasta maker, a mediocre Squash and Guitar player, and an enthusiast webscraping and MTG nerd.

Always happy to chat about research so please don’t hesitate to reach out!