The Golub Capital Social Impact Lab has an opening for a 2 year position as a post-doctoral fellow, working Professor Susan Athey and other faculty affiliates of the lab. The position involves working on a set of projects revolving around educational technology and serving as project manager and collaborator on research projects, as a coauthor on some projects and as an assisting researcher on others. It also requires interfacing with industry partners and supervising research assistants at various levels of seniority.
The ideal candidate is either preparing for an industry position, for example in a technology company, or an academic position in a field closely aligned with the lab, for which collaboration on the lab’s projects would serve as strong preparation. This position does not incorporate independent research by the fellow outside the scope of the lab; any independent research would be conducted outside of regular work hours and should be managed so as to not present a conflict of commitment to the lab.
Depending on the fellow’s skills and interests, the fellowship will create the opportunity to: create novel experimental designs, including adaptive and dynamic treatment regimes, bandits, and contextual bandits; run the experiments in collaboration with technology firms or on tech firm platforms; use and develop cutting edge methodology for working with large data sets, using university infrastructure or the infrastructure of tech firms; including tools of machine learning and causal inference; develop coding expertise for publicly released software; and/or develop expertise in managing large-scale empirical projects with large code bases written by teams.
The position involves working on 2-3 project areas at the time. Following are broad contribution areas:
-
Application areas:
-
Consumer behavior around digital learning technologies
- Providing information to low-income workers about job options and potential related educational programs
- Developing low-cost assessments of learning achieved through educational technologies
- Measuring the welfare effects of potential new interventions and technologies
-
-
Methodologies:
- Methods from behavioral economics and consumer behavior
-
Analytics to support proposals for improvements in application design
-
Developing, evaluating, and implementing personalization, e.g. personalized recommendation systems
- Analyses may include the following methods depending on the research question and state of the project:
- Descriptive analyses involving aggregating and describing consumer behavior
- Analyzing natural experiments and field experiments
- Online experiments such as applications of bandits or improving upon A/B testing methodologies
- Adapting machine learning and causal inference techniques to be useful in the improvement of innovations for learning
- Simulations to explore properties of more complex, adaptive experimental design
-
Analyses and coding:
-
As a supervisor:
-
Developing an outline and workplan for each segment of work
-
Making architectural decisions about how the code will be structured to carry out the previous tasks
-
Supervising teams of research assistants in assembling data from a variety of sources, running analyses as well as visualizing results and creating tables in a way that is replicable and well-documented
-
Maintaining a public GitHub repository with publicly available datasets or code to read and process publicly available datasets, together with sample code to run newly developed machine learning methods on those datasets
-
-
As a coder:
-
Need to be quick and nimble with conducting data analyses and adapting code from tutorials and other projects to get answers and make decisions
-
Writing code and conducting analyses using a variety of modeling techniques from econometrics, statistics, and machine learning
-
Writing and editing highly optimized low-level code to estimate models with many latent variables using methods such as variational inference and stochastic gradient descent
-
Developing and maintaining open source software, primarily written in R or Python, and sometimes C++, and releasing it on GitHub
-
Conducting simulations and applications of these methods to publicly available or newly created datasets
-
-
Community building and project management:
- Recruit, train and manage student research assistants. This includes ensuring continuity over projects on which multiple research assistants might work on across time.
- Plan, organize, and oversee research projects. This includes identifying promising research areas, ensuring good progress on projects, and securing all necessary data, research assistants and software to conduct research.
- Contribute to website content and course material that relates to research projects. This includes writing about work to engage advanced student, funder and partner audiences.
- Manage relationships with external (non-academic) partners such as tech companies and nonprofits. This includes coordinating meetings and workflow with external partners, and selecting external partners and projects to collaborate on.
- Contribute to funding applications and IRB approvals.
The strongest applicants will have a variety of skills and preparation, and will have a strong desire to rapidly obtain any skills and experience that are lacking. Desirable skills and experience include:
- PhD in economics, business or management disciplines, statistics, or related fields, with deep expertise relevant to at least some of the problems and methods described above
- Excellent coding skills in multiple languages (R, Python, C++ are three we use regularly)
- Experience with large datasets
- Experience with designing, implementing and analyzing experiments
- Experience using cloud computing services such as AWS, Azure, etc.
- Experience working with Github or similar tools for collaboration and version control
- Experience supervising research assistants
- Strong writing skills: ability to create clear, logical, and concise writing; ability to write clear and unambiguous interpretations of graphs and figures; ability to explain research to broader audiences
- Strong communication skills: high-bandwidth communication, the ability to create useful notes and workplans from meetings, conduct meetings with partners in a professional manner, provide and receive peer feedback
- Curriculum vitae
- Job market paper
- Additional paper (optional)
- Additional paper (optional)
- Other material (in one file) (optional)
- Cover letter
- Video presentation of job market paper (optional)
- Letters of reference: 3