Megagon Labs Summer 2020 Internship Experience

Megagon Labs has six students working with us for the summer 2020 internship program. This summer is a bit unusual. Our interns have to work remotely under the pandemic. Here, we invited them to discuss their experience, ranging from their path to Megagon Labs, lessons learned, their favorite part of the internship, and more. 

Shabnam Behzad

University: Georgetown University, Ph.D. Program

What is your project? 

I was working on better understanding the content of the reviews, more specifically, we wanted to be able to find proper conclusions based on different sentences in reviews.  (Project team members: Behzad Golshan, Nikita Bhutani, Xiaolan Wang)

How did you make the path to Megagon Labs? 

I’m a Ph.D. student and my area of interest is NLP. In the past, I read a couple of publications from researchers in Megagon Labs, and the work they did was really interesting to me. So when I was applying for internships and noticed that Megagon Labs has open positions, I applied and I was fortunate to have the opportunity to work here during the summer.

What’s been your favorite part of the internship?

Meeting and working with the amazing people at Megagon Labs! Everyone is so knowledgeable, so willing to help and they are always responsive. I had really great mentors and I really liked how we worked as a team on the project.

Dongjin Choi

University: Georgia Institute of Technology, Ph.D. Program

What is your project? 

In my internship project, I built an interactive system for data programming of span-level tagging problems. This system aims to facilitate the labeling function building process by allowing users to interactively build labeling functions. (Project team members: Çağatay Demiralp, Sara Evensen)

What’s something you learned during your internship that you’ll take with you?

I learned two specific skills that I’ll take with my research journey after the internship. The first skill I learned is setting weekly and daily goals and keeping track of them. By doing that I was able to check my achievement and stay productive. The second skill I learned is using React.js for web development. This was especially useful for building interactive UIs. I would definitely utilize this skill in my future research.

What’s been your favorite part of the internship?

My favorite part has been working as a team. Although I couldn’t actually meet and work in-person because of the pandemic, I had no problem working with my mentors Cagatay Demiralp and Sara Evensen. They always tried to help me and were kind to me. Thanks to them, I was able to achieve more than I could do on my own. It was very enjoyable and rewarding to collaborate with the world’s best researchers in this field.

Akhil Dixit

University: UC Santa Cruz, Ph.D. Program 

What is your project? As a Research Scientist Intern at Megagon Labs, I am working on building a semantic-aware “minion” for our question-answering assistant. This minion is aimed at extracting semantic information from the questions asked by users and utilizing it effectively to search meaningful answers in the customer reviews corpus. (Project team members: Nikita Bhutani, Behzad Golshan, Estevam Hruschka, Wataru Hirota)

How did you make the path to Megagon Labs?

Since I study logic and databases at UC Santa Cruz, I was no stranger to the work of our ex-professor Wang-Chiew Tan, who now leads the research efforts at Megagon Labs. In January 2020, when I was actively looking for summer research internship opportunities, I came across a post in the DBWorld mailing list by Wang-Chiew about an internship opening at Megagon Labs. Although it was inclined towards NLP and AI which aren’t particularly my areas of research, I was excited to apply as I saw a wonderful chance for me to get out of my comfort zone and acquire new skills. A couple of weeks later, I was interviewed by Behzad Golshan (who’s one of my mentors now), and a few more weeks later, I received my internship offer! 

What’s been your favorite part of the internship?

My favorite part of the internship is the flexibility and freedom I have here at Megagon Labs. My previous internships/jobs in the industry have mostly been in software engineering, and this is the first time I’m working as a research scientist intern. Here, I have the freedom not only to identify problems to work on but also to explore and experiment with multiple approaches to solve them. This is both fun and challenging at the same time and I’m enjoying it a lot. Moreover, I’m part of an amazing team that respects and encourages my ideas and motivates me to bring out the best of what I can do during my stay here. I hope to meet my team in-person soon!

Peter Griggs

University: Massachusetts Institute of Technology

What is your project? I’m working on a system that improves the user experience of doing exploratory text analytics, specifically on a part that is an extensible algebra for text analytics and visualization that powers the coordination between parts of the system. (Project team members: Çağatay Demiralp, Sajjadur Rahman)

How did you make the path to Megagon Labs?

I’m an Meng student in the Data Systems Group at MIT. I saw an email from Çağatay Demiralp about the internship opportunity on our group’s mailing list. After talking with Çağatay, I found that I had several shared research interests with Megagon Labs. I’m doing my masters thesis on data processing for data visualization systems and am interested in learning more about text analytics, so it was a perfect fit!

What’s something you learned during your internship that you’ll take with you?

This was the first time that I’ve designed and implemented a non-trivial system from the ground up, so I learned a lot about developing system design and prototyping. I think that my mentors did a really good job of helping me to define concrete steps that I could achieve along the way, which made everything I did feel less intimidating. Overall, I’ve had a great experience interning with Megagon Labs and it has worked out really well even though it was remote!

Chieh-Yang Huang

University: Penn State University, Ph.D. Program

What is your project? Salient facts are sentences that are informative, concrete, and unique. I worked on a project to extract salient facts from large amounts of reviews automatically, by applying an unsupervised method. (Project team members: Jinfeng Li, Nikita Bhutani, Alexander Whedon, Yoshi Suhara)

What’s something you learned during your internship that you’ll take with you?

One thing I remember and I would like to take with me is to take actions carefully and step by step. Before joining Megagon Labs, for most of the project, I would first survey some related works, come up with the idea, design and implement the algorithm, and then conduct experiments and analysis. If the algorithm doesn’t work, I will go back to the step of thinking the idea and do all the process again. However, my mentors always want me to conduct small and simple experiments to confirm that the idea is reasonable and likely to work before putting a lot of effort into the algorithm design and implementation. Though this might slow down the process a little bit, it actually helps us to finish the whole project quickly. I would try to follow this in my future research projects.

What’s been your favorite part of the internship?

It is hard to say if I like it or do not like it, but the most different thing about this internship experience is the remote-working environment. I really like the flexibility of remote-working where I can basically plan my schedule by myself. Staying at home, I can have dessert anytime if I want. However, remote-working prevents us from knowing more about other people in Megagon Labs. One important thing about the internship is to make connections with lab members and interns from different universities. Sadly, it does not happen in such a weird situation. Overall, though it is hard to describe this experience, I think I like it more.

Zhengjie Miao

University: Duke University, Ph.D. Program

What is your project? I am working on discovering data augmentation policies for NLP tasks following my last internship project. This work can also be extended to non-NLP tasks including entity matching and data cleaning. The purpose of my project is to reduce the label requirement by wisely leveraging augmented text data. (Project team members: Yuliang Li, Xiaolan Wang)

How did you make the path to Megagon Labs?

I had a very nice experience last summer. Both my mentors and I are still interested in the topic, and we continued collaborating in preparing the Snippext paper. During the collaboration we agreed that I would return to Megagon Labs to keep working on this project.

What’s been your favorite part of the internship?

My favorite time last summer was the coffee time that we made pour-over coffee together! Unfortunately this year the internship is remote:(. Besides this, I really enjoyed the research atmosphere. There are talks and very helpful comments, and my mentors are creative and supportive — not only in the project, but also they care about my research career a lot.

If you would like to learn more about our internship program, or to apply, check out our internship page. 

Megagon Labs summer 2020 interns zoom discussion

Written by our summer interns and Megagon Labs


More Blog Posts: