Hi, I'm Aiden.
I work with AI
Currently a Solutions Architect at NVIDIA, working on world models, autonomous vehicles and robot learning. I enjoy building intelligent systems that solve real-world problems.
I've set aside one 30-minute session each week for a quick chat. Feel free to book a call with me!
About Me
NVIDIA
Solutions Architect
U.S. Army DEVCOM Army Research Laboratory
Research Fellow (AEOP Fellowship)
USC × U.S. Space Force
Project Lead, rocket launch detection
KAIST
Visiting Student Researcher
Timonier Systems
AI/ML Engineer
University of Southern California
M.S. Computer Science (AI specialization) · 3.95 GPA
Carleton College
B.A. Computer Science · 3.63 GPA
Sophia University, Tokyo
Language Exchange Program
Contact
Interests
Long-term, I hope to advance autonomous exploration for high-impact missions in space!
Currently, I build world models for autonomous vehicles and robots that ground prediction in interpretable signals, and I want to fuse sensing beyond human perception so robots can act reliably where cameras fall short, from off-road terrain to space.
Honors
- Sigma Xi, The Scientific Research Honor Society2025
- Finalist, $100K Min Family Engineering Prize (top 3 of 86 teams)2025
- AEOP Research Fellowship, U.S. Army Research Laboratory (<8% acceptance)2024
- NeurIPS 2025 Reviewer2025
- 2× All-Academic Athlete, NCAA Division III2023
My Stack
Book a 1 on 1 chat
One 30-minute session each week, open to anyone.
Selected Works
Aha!
Predicting What Matters Next: online highlight detection without looking ahead.
Rocket Launch Detection
Predicted stealth rocket launches from satellite imagery and NOTAMs, catching a real launch three days before ignition. Top 3 of 86 teams in the $100K MEPC.
Robotics: Learning to Walk
Building a robot that learns to walk from scratch, simulation to reality in 100 days.
SELDVisualSynth
Generating diverse audio-visual 360° soundscapes for sound event localization and detection.
2D3MF
Deepfake detection using multi-modal middle fusion, with state-of-the-art results.
Long-Tailed Biases in MMLLMs
Investigating long-tailed biases in multi-modal large language models.
ARS
Graph neural networks for Amazon product rating link regression, achieving 1.01 RMSE.
ALLM
An auditory large language model for embodied agent perception.
GPT-3 From Scratch
Building GPT-3 from scratch using PyTorch.
Stable Diffusion From Scratch
An open-source implementation of the Stable Diffusion model.
Research
Publications & Patents
- ConferenceNeurIPS 2025Aha! Predicting What Matters Next: Online Highlight Detection Without Looking AheadAiden Chang, Celso de Melo, Stephanie LukinAdvances in Neural Information Processing Systems, with DEVCOM Army Research Laboratory
- ConferenceIEEE Haptics 2026Language-Guided Multimodal Texture Authoring via Generative ModelsWanli Qian, Aiden Chang, S. Lu, M. Gu, Heather CulbertsonIEEE Haptics Symposium
- WorkshopDCASE 2025Generating Diverse Audio-Visual 360° Soundscapes for Sound Event Localization and DetectionA. S. Roman, Aiden Chang, G. Meza, I. R. RomanDCASE Workshop, with Sony and Queen Mary University of London
- WorkshopNeurIPS 2025Community-Aligned Behavior Under Uncertainty: Evidence of Epistemic Stance Transfer in LLMsPatrick Gerard, Aiden Chang, Svitlana VolkovaNeurIPS 2025 Workshop: Epistemic Intelligence in Machine Learning
- PatentFiled Feb 2026Cellular Network Operations Reasoning ModelsAiden Chang, A. Canaveras, A. UskudarU.S. patent application, NVIDIA
- ArticleFeb 2026Building Telco Reasoning Models for Autonomous Networks with NVIDIA NeMoAiden Chang, A. Canaveras, A. Uskudar, A. PhadkeNVIDIA Technical Blog
- ArticleDec 2025How to Scale Data Generation for Physical AI with the NVIDIA Cosmos CookbookP. Mishra, Aiden Chang, E. LlontopNVIDIA Technical Blog
Talks & Workshops
- Guest lectureSep 2026How to Build End-to-End Physical AI Systems for RobotsUSC CSCI 566, Deep Learning and its Applications · Los Angeles
- WorkshopApr 2026NVIDIA & Marlowe: Training Robots with World Foundation ModelsStanford HAI & Marlowe, co-presenter · Stanford
- WorkshopApr 2026Accelerate Robot Learning with NVIDIA Isaac Lab and NewtonGoogle Cloud Next 2026, creator & lead instructor, 80+ participants · Las Vegas
- LabMar 2026Fine-Tune a Telco Reasoning Model: Synthetic Data, Tool Calling, and EvaluationNVIDIA GTC 2026, creator & lead instructor · San Jose
- PosterDec 2025Aha! Predicting What Matters NextNeurIPS 2025 · San Diego
- PosterDec 2025Language-Guided Multimodal Texture Authoring via Generative ModelsIEEE Haptics Symposium · Reno
Hobbies
College Football
Carleton starting running back, three-year starter
Nationals in Colorado
Northstar Ski Team, sponsored by The North Face, ranked 47th in the nation
Coaching for fun!
Robotics
Cross Country Skiing
Skiing
Rugby
HS Football
My Blog
Paper accepted to NeurIPS
My paper, Aha!, was accepted to NeurIPS 2025!
VLM from Scratch
Building a multimodal vision-language model from scratch.
Joining NVIDIA
Woohoo! Joining NVIDIA as a Solutions Architect to work on AI projects!
Starting at the U.S. Army Research Laboratory!
Becoming a full-time AI/ML researcher, working with robotics.
Beginning at the DI Laboratory @ KAIST!
Researching ethical AI under Professor Whang at KAIST's DI Lab.
SDA TapLab Presentation
Presented our work on detecting rocket launches using AI to the U.S. Space Force.
ARS
Building Amazon recommendation systems with graph neural networks.