About Me
Hi, I’m Anirudh. I am a final-year at MIT pursuing my B.S. in Computer Science and Engineering (Course 6-3). I am also concurrently pursuing my MEng concentrated in Artificial Intelligence through the EECS department. I’m interested in harnessing technology and machine intelligence to streamline the systems in our own lives. Recently, I worked as a machine learning engineer at Moveworks, enabling user-level personalization for the company's agentic AI chatbot for employee support automation. Previously, I worked as an embedded software engineer on Nuro’s self-driving vehicles.
I am an eager learner fascinated by the prospect of harnessing machine learning to devise systems capable of replicating human intelligence. People operate in a physical world, and I have been interested in building useful scene representations for robots through my research with MIT's Scene Representation Group at the Computer Science & AI Lab (CSAIL).
Some hobbies that I have include making Youtube videos, teaching high-schoolers through Momentum AI, learning languages (I speak 5!) and dancing for MIT’s competitive fusion dance team, MIT Mirchi.
What am I up to?
Exploring the frontiers of robotic intelligence and AI perception as an undergraduate researcher with the Scene Representation group.
I enjoy building software products that solve everyday problems. Check out some of my favorites!
Shedding light on my tech journey at MIT and beyond through videos. Built an audience with over 465,000 views and 8,400+ subscribers.
AI & Robotics Research
I am fascinated by the prospect of enabling generalizable everyday automation (think personal robots or self-driving cars!) using recent advances in deep learning.
In my work as a SuperUROP scholar in the Scene Representation Group with Boyuan Chen and Dr. Vincent Sitzmann, I investigate novel neural scene representation approaches to push the boundaries of robotic manipulation. I also collaborate heavily with the Robot Locomotion Group and professor Russ Tedrake.
I also develop curriculum for Momentum AI, an MIT organization that hosts camps for students in the Boston area to learn about artificial intelligence.
My Robotics Research Projects
Grasp Inpainting: An End-to-End Approach
Advances in Computer Vision Final Project
In this research project, I developed an end-to-end dataset generation pipeline using inpainting techniques to supervise downstream grasping tasks through imitation learning. My method leverages recent advancements in image generative modeling, off-the-shelf segmentation, and single-view 3D reconstruction to inpaint robot hands on any image, enabling the self-supervision of imitation learning tasks.
Intersection Data Imputation
TRB (Transportation Research Board) 2023
This project involved developing environment models describing 9 large American cities for downstream reinforcement learning tasks, specifically to train an eco-friendly autonomous vehicle fleet control policy. Our study analyzed hundreds of thousands of intersections on OpenStreetMaps and found several examples of intersections without a clear representation of intersection geometry. Our method, which was developed in collaboration with the Utah Department of Transportation, was able to recover intersection geometry, turn data, speed limit, and gradient information for use in the lab's reinforcement learning projects.
Learning Dexterous Robot Grasps from Human Video Demonstrations
MIT SuperUROP Project
As a part of MIT's SuperUROP program, this proposed project plans to explore methods to learn dexterous grasping skills solely through human video demonstrations. We plan to solve this problem with the key challenges of mobile manipulation in mind, where a robot is limited to its on-board perception capabilities to complete tasks in unknown environments. We hope to leverage recent advancements in 3D perception and scene representation to learn a functional and accurate world model to help a robot build a dynamic understanding of its surroundings.
Software I’ve Built
Youtube
On my Youtube Channel, I provide an inside look into what life is like at MIT, complete with vlogs, interviews, and Q&As to take my viewers through my journey in tech. I hope to bring the incredible conversations and experiences that I have been fortunate to experience at MIT and inspire my viewers to pursue their dreams in tech. Check out some of my vids!
An Exciting Day In The Life of an MIT Computer Science Student
An exciting day of my life during midterms, complete with beautiful shots of campus, sailing on the Charles River, an insider's look at my dance team, and tips and tricks about surviving at the #1 STEM school in the country!
An Inside Look at San Francisco's Cutting-Edge AI Hacking Scene! | Scale AI Hackathon 2023
Watch me document my exhilarating day at the Scale Generative AI hackathon, where I show our roller-coaster experience developing SnapLingua, an app to learn vocabulary in your downtime.