GRASP Lab 2026 Summer Internship Round-Up

Amazon Robotics. It was an in-person internship in Westborough, Massachusetts, so I relocated there for the duration of the internship.

Amazon. Outside of Philadelphia. I had to relocate to California.

Meta, Location: Redmond, WA (onsite)

Roblox, in San Mateo, California (Hillsdale). The internship was on-site.

The company I interned with this summer was Summer Robotics, Inc. The internship was based in Campbell, California, outside of Philadelphia, and was fully on-site. As a result, I had to relocate to California for the duration of the internship.

Built Robotics, local in Philly, no relocation.

The entire interview and hiring process took about one week, and I received my offer around December 2025. The onboarding process was straightforward, with the necessary paperwork and training completed before and at the start of the internship.

I saw the internship online on their website. I applied online, and they reached out for a few rounds of interviews.

My advisor gave a talk at the company where my paper was featured (PhysGS: Bayesian-Inferred Gaussian Splatting for Physical Property Estimation), and this led to an interview with the team at Meta, as my project was related to the work that the team wanted to investigate.

I first heard about the internship through my professor’s recommendation. I had previously been hired by a startup, but by the time I joined, the company had been acquired by Roblox, so the application process should be different from Roblox’s usual process. We just had a single Zoom interview, and that was all.

I first learned about Summer Robotics, Inc. through a Google search for companies working in perception and robotics. Their work aligned closely with the courses and projects I had completed during my first year at UPenn, so I reached out directly via email with my resume. After completing the interview process and being selected, I was assigned a mentor who introduced me to the company’s technology and gave me initial tasks that helped me quickly get up to speed.

The company had a collaboration with the research group that I’m a part of at Penn (xLAB). Since I was referred through the research group, the application and onboarding were super smooth.

I applied to a few other internship opportunities. I chose Amazon Robotics because it was the only offer I received, and, more importantly, the position aligned very closely with the type of robotics work I wanted to do (more on the R&D side).

I applied to Honda Research Institute, Qualcomm, Google, and Meta. I ultimately chose to pursue the Meta internship because they were making an extremely interesting bet in robotics, particularly in dexterous manipulation, and I felt the work there was the best fit for a 16-week-long internship.

I received an interview opportunity from Adobe, and I could have chosen between the startup and Adobe. I chose the startup because Adobe’s opportunity was more focused on audio, while my work and my lab’s research focus more on images and video.

I applied to around 5-6 internship opportunities. I chose this internship because the company’s work aligned closely with the courses I had taken and the research I was working on at GRASP Lab, making it a strong fit for my academic and career interests.

I applied to about 200 in total. The project I got for this one was hands down the best and most relevant to my career goals.

Yes, I am an international student and participated in UPenn’s CPT process. The CPT process was very smooth and straightforward. It took me about two weeks to get it approved (10 business days).

Yes. It is easy. Make sure to contact OAS and ISSS and just follow the instructions on their website. Make sure you start early, as there might be some wait time.

I am an international student and used CPT. Make sure to take at least 6 CUs during your first year, and use CIS 9999 as much as possible. I took seven classes but did not complete 6 CUs, because TA training does not count toward CUs, so be careful. Other than that, the process was smooth. Also note that the CPT application deadline is two weeks before your job start date, and processing usually takes two weeks, so you should start at least four weeks before your job starts.

Yes to the first three questions. The CPT process was easy to navigate. I just followed the instructions available on the website: https://grad.engineering.upenn.edu/student-handbook/academic-options/curricular-practical-training/.

Yes, and I did do CPT. It was very easy to navigate. Advice: start early (as early as the start of the academic year).

I would recommend applying to positions that closely match your skills and research interests, and reaching out directly to recruiters whenever possible. I would rather apply to five positions where I have relevant connections or referrals than 50 positions with no connections. I also strongly recommend joining a lab at UPenn, as the research experience and connections can be extremely valuable. During my Amazon interview, we spent about 15 minutes discussing GRASP and the projects I had worked on in the lab, which gave me a great opportunity to connect my research experience to the position.

Apply to jobs that match your experience and interests. It is hard to get a position that is not reflected by your experience.

Getting a referral from someone working in the company versus cold-applying may be beneficial to get more positive responses. Further, on your resume you should feature the research or industry experience and projects you have been a part of, or led, preferably.

Maybe ask your professor.

From an application perspective: I would recommend starting early and applying broadly, while still focusing on opportunities relevant to your interests. Visit each company’s website and learn about their technology to gauge whether their work aligns with your research and career goals.

From a resume perspective: keep your resume updated. Highlight relevant research, coursework, and project experience in both your resume and cover letter to demonstrate how your background aligns with the role.

Start your applications as early as possible and be regular with them. Try getting referrals, as those significantly boost your chances. Discuss with your peer network – you’ll often find someone who’s already interviewed for the role you’re looking at, and they could be of immense help.

I worked on robotics simulation using MuJoCo and Isaac Sim, as well as mechanical design and controls. My day-to-day work consisted of attending meetings, working on technical tasks to advance my assigned project, and collaborating with my team to meet project deadlines. I also occasionally presented my work to the broader team to provide updates on my progress. Overall, the internship involved a combination of simulation, mechanical design, controls, and robotics development.

I do hardware and machine learning.

Dexterous manipulation.

Research on efficient video generation.

My responsibilities spanned robotic perception, machine learning, 3D sensing, and data generation. I worked with laser-event sensing technology to generate point clouds and annotated registration datasets for geometrically challenging industrial parts. I also developed methods for automatically determining registration tolerances using PCA and second-moment-of-inertia analysis, explored AI model interpretability using Integrated Gradients, and worked on sensor calibration and sensing configurations for capturing fine geometric details.

My project was on the end-to-end development of an autonomous ground robot for high-fidelity solar site surveys. Daily responsibilities included almost all aspects of robotics, including hardware design, assembly, BOM management, software design and testing, and field testing.

My internship is a seven-month co-op. I am expected to work 40 hours per week, in person. The longer duration has allowed me to become deeply involved in my project and contribute to it throughout multiple stages of development.

I had no set hours. I had update meetings to show the progress of the research and get actionable feedback from peers. My internship lasted for 16 weeks.

I worked full-time from July 14 through August 21, then part-time from August 25 through December 17.

The duration of my internship was approximately two months, and my working hours were from 9:30 am to 5:30 pm.

Full-time for about three months. As a remote team, we structured the work around daily stand-ups and milestone-based check-ins, with the run-up to our product demo being the busiest period.

Absolutely. GRASP was a great preparation for my internship. The labs at GRASP work on many of the same problems that companies in the robotics industry are trying to solve, so the research experience was very relevant. I also found several of my courses, such as MEAM 5170 and ESE 6500, to be very helpful. It ultimately depends on the type of internship you are pursuing, but if you know in advance what kind of position you will be doing, I highly recommend taking courses related to that field. It can make a significant difference.

Yes. GRASP Robotics MSE is the BEST program to get you ready to be a good engineer. Project classes and the projects in the research lab really prepared me well.

Yes, my time at GRASP helped prepare me for the internship, as in the projects that I am participating in have similar update meetings. This workflow is very similar to industry, although one difference is that you get a lot more feedback while in industry, helping interns speed up their workflow.

It might have been better if GRASP had emphasized that we need 6 CUs and that TA training does not count as a CU. Students should make sure to use CIS 9999 as much as possible.

Yes, my time at GRASP helped prepare me well for the internship. The coursework and projects I completed during my first year gave me a strong foundation in robotics, machine learning, perception, and working with 3D data.

Definitely, my coursework and exposure to facilities at GRASP were immensely helpful. In particular, I was directly applying my learning from the machine perception, design of mechatronics, learning in robotics, and roboracer courses to the development of the robot.

Definitely the lunch and ice cream breaks! Amazon has a huge food culture, and my colleagues and I definitely took advantage of all the free food whenever we could.

Interacting with my mentors and peers was the best part of the internship, as I learnt a great deal about the landscape of the field at lunches and other casual talks. I also got to explain my work, which helped me get valuable feedback and make impactful design decisions in my project.

Everything was good: a good environment for research and a good stipend—everything I needed.

My favorite part of the internship was the freedom to explore and getting the opportunity to work across different areas of robotics, from machine learning and robotic perception to hands-on work with sensing hardware. Also, the interesting career development and interview workshops led by senior team members with experience at companies such as Google and Samsung! As well as the fun team activities organized throughout the internship.

Getting to work end-to-end on the robot, from sensor integration through to perception and deployment, and then seeing it run in a real industrial environment rather than just in simulation.

Yes, absolutely. Besides the technical skills I developed, I learned a lot about working effectively within a team. I gained a better understanding of the importance of meetings, communication, and collaboration in a professional environment. I also feel that my leadership skills have grown throughout the internship, along with my technical skills in both hardware and software. These are skills that I will definitely carry into my future research and career.

The main skills I improved in my internship were adaptability and creating a high-bandwidth research loop. Firstly, I had to adapt to the tools and infrastructure available at Meta, which took a bit of time for all interns to get acclimated. Secondly, since I had a high density of meetings with my collaborators, I needed to make sure I had rapid updates from my project, and be able to communicate this effectively, as well as elicit the feedback I needed to make further progress or unblock myself.

I think I learned how to conduct research with limited compute, interestingly through the experience of having access to larger compute resources. I also developed an intuition about the effects of batch size. I believe the research will lead to a paper soon.

Yes, the internship strengthened my algorithmic acumen while also giving me valuable practical, hands-on experience working with robotic perception systems and sensing hardware. These skills and experiences will be highly valuable for my future studies at Penn and for pursuing a career in the robotics industry.

Yes, I believe the skills I gained have prepared me well for my 2nd year at Penn, where I’ll be focusing on applying advanced motion planning and control strategies for my master’s thesis.

Yes. This internship made me more interested in manipulation robotics and simulation. Working on these areas gave me a better understanding of the types of problems I enjoy solving and the direction I may want to pursue after graduation.

I was a research scientist intern, and this experience gave me further insight into what industry-grade experience looks like, and how it differs from the academic experience. This will definitely shape the decision on my post-grad plans.

Yes, absolutely! The internship reinforced my interest in pursuing roles that combine robotics, machine learning, and perception, while also giving me a better understanding of the type of work and environment I would like to pursue after graduation.

Yes, I think it’s given me a good overview of how fast-paced robotics startups work and which departments I would be interested in working in.

I met a lot of amazing people during my internship. I had the opportunity to work with students and engineers from CMU, MIT, UT Austin, Penn, and other universities. I also had the privilege of being mentored by engineers who previously worked at Google and other big tech companies.

Certainly, I made many connections which I hope to leverage past this summer, either as a returning intern in the future, or in some research collaboration.

I plan to continue working there as a part-time intern.

Yes, I made many valuable connections with team members who have experience at companies such as Google, PayPal, and Samsung.

Yes.

The internship followed a typical 9-to-5 schedule, so maintaining a good balance was relatively doable. I did not have many other major responsibilities during the internship, which allowed me to focus on my work during the day. However, I would have had some time for additional research.

I just did the work 🥲. You can do everything.

I balanced my internship and research responsibilities by working on my research at the GRASP Lab during weekends or after office hours.

The internship was quite hectic, so I was only focusing on that during the summer and did not do any additional research work on the side.

Yes, in a heartbeat. I would highly recommend it to future GRASPees. The networking and connections are a major benefit, and the internship also gives you a great perspective on how a large robotics company operates and how research and engineering are translated into real-world products.

Yes.

Yes, this was a great experience, and Meta is positioned to be a strong leader in the manipulation field, which requires knowledge and skills from all subfields of robotics. I would certainly recommend this to future GRASPees.

I recommend that GRASP students intern at Roblox. It is simply a wonderful place. I hope to work here full-time in the future.

Yes, I am very grateful for the opportunity to work closely with my manager and mentors and to receive guidance and support from other team members throughout the internship. It was a valuable learning experience, and I would definitely recommend it to future GRASPees.

Hands-down, yes.