Maximilian Du

"And still 24 hours, maybe 60 good years, it's really not that long a stay..." -JB

Hey there! I am a second-year Ph.D. student at Stanford University working with Shuran Song in the Robotics and Embodied AI (REAL) Lab. I'm broadly interested in how robots can learn and reason like humans and animals. I am grateful to be supported by the Knight-Hennessy Fellowship and the NSF Graduate Research Fellowship.

I completed my undergrad degree in computer science at Stanford, where I conducted research in Chelsea Finn's IRIS lab on imitation learning methods. In the months before my PhD, I interned at the Navy Marine Mammal Program, where I worked hands-on with dolphins and sea lions. This once-in-a-lifetime experience convinced me that a better understanding of animal cognition will be critical in crafting better AI algorithms.

Beyond the robot lab and my continued involvement with the Navy dolphins, I'm a writer and an oral historian. Sometimes this journey takes me to advocate for animal care professionals, their animals, and responsible zoological facilities. But most of the time, I enjoy writing short stories. I hope you can read them soon.

Email  / Google Scholar  /  GitHub  /  CV

Maximilian Du

News

June 2025: I became the first (part-time) employee of ACE, a company that develops video games for marine mammals and beyond.

April 2025: Honored to receive the NSF Graduate Research Fellowship!

March 2025: Our dolphin video game project CV-EVE won 2nd place at the International Marine Animal Trainer's Association (IMATA) conference.

December 2024: My first op-ed published about my oral history work.

Sept 2024: I started my CS Ph.D. at Stanford.

May 2024: My first narrative audio story published about whales!

June 2024: Finished undergrad and started my internship at the U.S. Navy Marine Mammal Program in San Diego.

March 2024: Honored to receive the Knight-Hennessy Fellowship!


Research

I'm interested in robot learning, cognitive science, and behavior. Most of my current research is about making robots more adaptive by incorporating memory, curiosity, and plasticity. I also work on making robots that interact positively with people.


DynaGuide: Steering Diffusion Polices with Active Dynamic Guidance

Maximilian Du, Shuran Song

Submitted to Neurips 2025

Website  /  Paper  /  Code

Robot policies are getting more complicated, which also means that tuning them also becomes tedious. In DynaGuide, we propose a novel way of steering policies by using a dynamics model to influence the action inference process of a robot policy. DynaGuide works on any diffusion policy, including off-the-shelf real robot policies.

To Err is Robotic: Rapid Value-Based Trial-and-Error during Deployment

Maximilian Du, Sasha Khazatsky, Tobias Gerstenberg, Chelsea Finn

Website  /  Paper

When deploying robot policies, we need to know when we've made mistakes so we can recover from them and try something else. In this paper, we propose a mistake detection approach that uses the Bellman error of a learned value function. This simple trick boosts success rates of trained policies, all without modifying the base policy.

Behavior Retrieval: Few-Shot Imitation Learning by Querying Unlabeled Datasets

Maximilian Du, Suraj Nair, Dorsa Sadigh, Chelsea Finn

Robotics: Science and Systems 2023

Website  /  Paper  /  Code

When training robots on large datasets, we need to select the right data to improve its performance without confusing the robot. In this work, we propose a simple data filtering approach using distances in a latent representation. This data filter has large impacts on robot performance in simulation and on a real setup.

Play it by Ear: Learning Skills amidst Occlusion through Audio-Visual Imitation Learning

Maximilian Du*, Olivia Lee*, Suraj Nair, Chelsea Finn

Robotics: Science and Systems 2022

Website  /  Paper  /  Code

Operating in the real world, sound cues can tell us information even when vision cues fail, like rooting around a dark bag. In this work, we add the sound modality to a robot by adding a microphone to the gripper. We show that it is able to locate and extract hidden keys from a bag.


Writing

I work primarily in short fiction and creative non-fiction, although I have produced in other mediums, including audio podcasting and journalism. A lot of my creative inspiration comes from my marine mammal trainer friends, but I'll write about anything that looks at our complicated relationship with ourselves, our past, and the world around us. A small selection is work is below. For a more extensive list, click here.


Sharks, Palominos

Jaden is a playboy who takes breaks from promiscuity to visit his divorced father, dying from cancer. Kyrie is a killer whale trainer who keeps returning to free drinks and dumb men. It's not gonna be a romance story. Nothing lasts beyond a blink of time. But what can we gain from something we know we'll lose? This is my current short story and I hope to finish it by Fall 2025.

All the Stars in the Air

Preprint Link

Growing up, my immigrant mom wanted me to blend in with the American kids who had deep roots. She bribed me with fireworks and other things that explode. Winner of a Stanford creative writing prize.

Orca Boy

Podcast Link

A long time ago, my childhood friends called me Orca Boy. But when a SeaWorld trainer named Dawn was killed by an orca, my love for whales turned to shame. It all changed when I met Dawn's best friend, a whale trainer named Lyndsey, who led me back to SeaWorld on a journey of reclaiming the Orca Boy that I thought was gone. An immersive audio story produced with the Stanford Storytelling Project.

Chess with the Beasts: An Oral History Collection

Book Page

In this oral history book, I combine the personal, academic, and literary angles to tell a new story of the human-animal relationship through the individual lives of exotic animal trainers. I look at how our relationship to animals has changed and how we should look at animals in generations to come. This ongoing project started in 2020 and it has taken me to many amazing people and places.


Misc Projects


Course Notes

Access My Notes

During my undergrad, I took hundreds of pages of notes on topics that range from probability to psychology. They cover all levels of knowledge, from undergraduate introduction to cutting-edge AI algorithms. I'm working on making all of these notes public, and I hope that they can be helpful to some people. Currently they're hosted as raw HTML but I'm in the process of turning them into an interactive page.

Dolphin Computer Vision Game Controller

During my time with the Navy Marine Mammal Program (and now at ACE), I developed a computer vision controller that allows dolphins to interact with a video game system through their body motions. If you are a marine mammal facility and believe that your animals could benefit from a similar system, please reach out to me or Kelley Winship.

Media Annotator

Code

This simple Python-based program allows you to use keyboard shortcuts to annotate audio, video, and live events with timestamped comments. The program will export your annotations to copy-and-paste text that you can add to any literature review notes. I rely heavily on this tool to review hours of videos for my book.