Meta Research Scientist Faced Challenges in Grad School

Name: Ronshee Chawla (he/him)

PhD: Statistics and Machine Learning, University of Texas at Austin, 2024

What was your main area of research?

My PhD dissertation focused on the design and analysis of decentralized machine learning algorithms for multi-agent systems. The goal of my research was to utilize collaboration in multi-agent systems for faster learning and better performance guarantees under realistic communication constraints.

What is your current job?

I am a Research Scientist for Meta in San Francisco Bay Area, California.

I am working on large scale machine learning problems to improve their experimentation framework.

My favorite thing about the job is that I have the opportunity to leverage the knowledge and skills I gained during my grad school to solve impactful problems at scale.

How did you find this position? What were the career steps you took to get to where you are now?

I reached out to the Director of the Adaptive Experimentation Group in Meta's Central Applied Science organization because their work closely aligns with my research interests and the focus of my dissertation. Although my skills didn’t perfectly match the requirements of that team, my profile was forwarded to the Ads Online Experimentation Group, where I am currently working.

PhD graduate ➡️ Research Scientist

Why did you decide to not pursue a career in academia?

My graduate school journey began in Fall 2016 at Caltech’s Electrical Engineering (EE) department, focusing on network information theory. By the start of the 2017 winter quarter, I realized that this field had become saturated, and I didn’t see much room for growth. I decided to pivot and focus on fundamental problems in statistics and machine learning. Despite excelling in the EE PhD quals, I found myself struggling with limited support from both the department and the graduate student office, which was frustrating.

Transferring labs wasn’t easy, as many professors I spoke to didn’t have funding to take me on as a student. As a new international student, I was still navigating through the complexities of the U.S. academic system and worried about my funding running out. I quickly learned that professors had to constantly secure funding to keep their labs running, which was a source of stress for many, regardless of their success.

Fortunately, I had also received an offer in 2016 from the University of Texas at Austin’s Electrical and Computer Engineering department. After reaching out to my potential advisor, they were able to convince the admissions committee to reconsider me, and I made the transition in Fall 2017. I am incredibly grateful to my advisor for facilitating such a smooth transition.

The challenges I faced early on left a lasting impact, making me question whether academia was the right path for me. The idea of running a lab and potentially not being able to support students financially — whom I would treat like academic children — was particularly troubling. What ultimately pushed me away from academia, though, was the arms race I observed in machine learning, especially with the rapid developments in large language models. The constant pressure to publish and chase the latest trends didn’t align with my belief in long-term, meaningful contributions to science and technology. It also took a toll on my personal well-being.

Leaving academia was a difficult decision. I had always dreamed of being a faculty member, making a lasting impact, and empowering students through mentorship. I thoroughly enjoyed my experience as a teaching assistant, and my students were consistently happy with my approach. However, I realized that pursuing a career in academia no longer aligned with my personal and professional goals, and I decided to explore other avenues to make an impact.

What advice do you have for someone getting their PhD and looking to pursue a career outside of academia?

1. Do you have tips for interviewing for non-academic careers? How do you recommend folks package their PhD skills for other jobs (e.g. transferrable skills)?

To prepare for interviews in non-academic careers, I recommend reaching out to alumni from your program or professionals in your network who have successfully transitioned from a PhD to a non-academic role. Ask them about their current role, responsibilities, how they prepared for interviews, and what resources they found helpful.

If you already have upcoming interviews, don’t hesitate to ask the recruiter about the specific areas you should focus on for preparation. They will likely be happy to offer guidance or share resources to help you get ready.

When it comes to packaging your PhD skills for non-academic jobs, it is important to clearly communicate how your research experience translates to real-world problem-solving. For instance, if your work is theoretical, such as designing and analyzing algorithms, explain how your algorithms compare to existing solutions or how they push the boundaries of what is possible in terms of performance.

If your research is more quantitative or empirical, make sure to clearly define the problem you are addressing (and provide some practical context, if possible), summarize the approach or solution you used, and highlight the impact of your work—ideally with quantifiable results.

By framing your PhD work in terms of practical skills and measurable outcomes, you will make a stronger case for how you can add value in non-academic settings. If you could also demonstrate how your work and skills could be helpful in addressing the business needs of the organization that you would like to work for, that is a bonus.

2. How did you prepare for the job you currently have (e.g. specific extracurricular, programs, volunteer service, research skills, etc.)?

To prepare for my current role, I focused on developing strong algorithmic skills, which are essential for scientists in tech. While I was already well-versed in statistics and machine learning, my exposure to algorithms was limited to computer science courses I took in high school. To bridge that gap, I dedicated time during graduate school to self-study algorithms and data structures. I followed Prof. Tim Roughgarden’s video lectures and practiced coding problems extensively on Leetcode to refine my skills.

I also interned with Uber in the summer of 2022, which provided valuable industry experience. The internship gave me insights into how problem-solving works in a real-world context — how to balance technical rigor with the business context, and how to prioritize practical, effective solutions over theoretical optimal ones. This experience was helpful in demonstrating my awareness of the expectations of the job during my behavioral interview. For anyone transitioning from academia to industry, internships are an invaluable way to gain this kind of hands-on, practical knowledge.

3. How did you connect with those outside academia - those looking to hire but also just other folks pursuing non-academic careers? Tips for networking as well as community-building?

One of the strategies for connecting with those outside of academia that was most effective was identifying researchers in industry who were working on topics related to my own research. I reached out to them with a concise email, offering to contribute my skills and experience to help address challenges they were facing, while demonstrating how my expertise aligned with their needs. I believed that given their familiarity with the subject matter, they would have a deeper appreciation for my work. This approach resulted in positive responses from several of them, which helped build valuable connections and expand my network.