Internships are more than just a first step into the professional world—they’re a chance to learn, explore, and contribute to meaningful research. At Megagon Labs, our interns actively advance projects in NLP, LLMs, data AI, human-AI collaboration, and compound AI, while gaining hands-on experience working alongside our research teams. In volume one of this series, we asked former interns to reflect on their time at Megagon Labs, sharing candid insights about what they worked on, what they learned, and how the experience has shaped their paths moving forward.
Stay tuned for volume two, where four more interns share their experiences.
Chihiro Taguchi
University: University of Notre Dame
Program: PhD in Computer Science & Engineering
Project Description: Adaptation selection of documents for retrieval depending on the task, in essence flexible RAG.
Mentors: Seiji Maekawa & Nikita Bhutani
What was your experience like while working with the team at Megagon Labs?
People respect each other’s research. Sometimes, we can feel like our research isn’t significant enough. Here at Megagon Labs, we have a positive environment that motivates me to work more on my project and helps me discuss any doubts.
My mentor, Seiji, always gives me really constructive feedback on my research.
Was there a moment during your internship that made you feel especially proud?
The first month I was given a high-level idea of what Megagon needed from my project, but not much direction on an approach or method. It was an open question. This is usually how they run the internship—it’s a partnership between the intern and the mentors. It allows us as interns to have a lot of creative control when it comes to the approach we take.
After a month or so, I had a meeting with Seiji, and I had an idea for adaptive retrieval. I explained it, and he was intrigued and wanted to try it.
I was proud of my idea because it wasn’t prompted by anyone but me. Now, I’m proud to implement my own idea and proud that other people with research experience think it was a valuable and interesting approach. This helped my confidence as a researcher.
How has the Megagon team helped you broaden or deepen your research interests?
I came from traditional NLP-machine translation. Now, I’m more interested in what LLMs can and can’t do. I can combine this new knowledge with my PhD. Through Megagon, I gained hands-on experience with LLMs and running experiments. And now, RAG is a big research interest of mine.
What advice would you give to future interns?
First and foremost, enjoy yourself. Make sure you manage stress, whether it’s from the internship or your personal life. Reframe and refocus so this becomes a good experience. You have a lot of support from everyone at Megagon Labs, so do check-ins and talk to people.
Farima Fatahi Bayat
University: University of Michigan
Program: PhD in Computer Science & Engineering
Project Description: Reasoning hallucination of LLMs in mathematical problem solving when they are provided with access to a Python Interpreter.
Mentors: Pouya Pezeshkpour
What was your experience like while working with the team at Megagon Labs?
It has been a very joyful experience because I truly enjoyed my research topic and project. Overall, the internship really expanded my knowledge of reasoning. Since my focus was on factuality during my PhD, I was newly exposed to a lot of dataset methods. I was very happy with every positive result that I saw.
I also really enjoyed the professional aspect. The lab environment is very supportive and friendly. My mentor was always available to answer my questions, and we met almost every day.
What surprised you most about doing research in an industry lab compared to academia?
I liked being at a research lab more than at a company that is purely industry. I liked that I had the opportunity to come up with creative solutions. I felt like I could contribute to research at Megagon Labs more than I did at my previous internship. I would like to come back to academia in the future, and the experience at Megagon Labs is the bridge between industry and academia.
Was there a moment during your internship that made you feel especially proud?
I had a lot of moments because I was working in another domain previously, and I didn’t really think that my past knowledge would be transferable to this domain. I really enjoyed and felt proud of myself when I was able to find problems in some of the best language models. I felt like I could really contribute to the field.
How did feedback from your mentor shape your project?
We started with a high-level idea of the project, but Pouya told me I could explore other ideas. He gave me creative freedom. I received a lot of good feedback throughout the project, and my mentors, Pouya and Estevam (research lab lead), were very supportive. They were very encouraging of every little bit of progress that I made. They helped me move the project forward in a short amount of time. Because the PhD environment can be harsh or dry, this felt like a nice change.
What advice would you give to future interns?
Academia and industry are very different. Megagon is a more flexible and exciting research environment than a PhD program. You will have more support, so enjoy the process. Also, ask for a lot of feedback. If you don’t speak up, no one will know you need help.
What was your experience like while working with the team at Megagon Labs?
Everyone is very friendly, so it’s been good. The office is very research-oriented, and I really enjoyed all the support and like-mindedness. The mentors are very willing to help. When I ran into roadblocks or technical difficulties, I had help to find alternative tools or solutions.
What surprised you most about doing research in an industry lab compared to academia?
I interned at big companies in the industry. The difference is that here at Megagon, I actually know everyone, and we work closer together. This internship position is also more research-oriented. I was able to interact with different people in different research domains, and I learned a lot from that.
How has the Megagon team helped you broaden or deepen your research interests?
Before applying for the internship with Megagon, I decided to switch my research interest from computer systems to language models. My internship experience at Megagon solidified that decision and reshaped the direction of my PhD and future career.
What’s an activity you do to de-stress during your internship or PhD life?
One of my favorite ways to de-stress during my internship was hiking. The Bay Area has many beautiful trails, and I especially enjoyed exploring spots where I could catch the sunset. If you are looking for a great place to go, I recommend Sierra Vista in San Jose, which has amazing views!
What advice would you give to future interns?
The internship process at Megagon allows you to shape and define your own project. My advice for future interns is to come in with as clear an idea as possible, or at least a well-defined research question that can guide your work. Having that clarity from the start helps you stay focused and make good progress. And if your project is about improving an existing method, it’s helpful to establish a benchmark or baseline early, so you have something concrete to measure your progress.
Written by Megagon Labs
Megagon Labs conducts internships during Winter and Summer terms. Apply today!
