CV

Staff Machine Learning Engineer, Data Intelligence

XMotors.ai

Data Curation, Causal Representation Learning, Out-of-Distribution (OOD) Detection, Foundation Model

• Pioneered a causal embedding retrieval system for VLM-based autonomous driving, utilizing hidden states as decision-critical representations to perform root cause analysis on model failures; validated embedding effectiveness via Lipschitz continuity and prediction consistency checks
• Designed an automated “Data Debugging” instrument that correlates model failures with training data distribution; utilizes trajectory dispersion and instruction compliance metrics to automatically categorize failures into “Data Quality Issues” vs. “Model Capacity Issues,” accelerating the iteration loop for safety-critical interventions
• Built a high-throughput data processing & validation framework with Ray, integrating embedding-based similarity search and gradient-based influence analysis; established a standardized quality gate for post-training validation that prevents model regression caused by ambiguous or noisy data
• Developed a bicycle-model-based trajectory smoother to convert ego pose sequences into continuous control variables (acceleration, curvature), improving RL policy alignment and stability during deployment
• Enhanced lane-centering and speed control by deploying CiLQR-based trajectory generation for overseas OTA releases; restructured TJA (Traffic Jam Assist) to improve low-speed following and intersection handling, significantly reducing user complaints in Europe
• Implemented core planning and control modules including traffic light handling logic and intersection strategies, reducing red-light violation counts from 0.6 to below 0.2 per 100 km

Senior Robotics Engineer, Product & Core Dev

Dexterity Inc.

Robot Control, Motion Planning

• Designed and optimized motion planning and control frameworks in C++ for high-performance robotic systems
• Implemented end-to-end trajectory tracking in C++, from velocity profiling to driver control, achieving fast, smooth, and precise robot movements
• Developed robust collision detection and avoidance algorithms in C++ to ensure safe robotic navigation
• Built custom simulation toolchains with integrated physics engines to replace ROS-based environments for robotic testing and development
• Designed and deployed a Python-based line-kitting application for a major client, defining workflow logic and integrating it into operational environments
• Integrated perception models, control logic, and sensor modules to deliver full-stack robotic applications

Master of Science

Stanford University

Computer Science, 3.9/4

• Specialization in Articifial Intelligence

Bachelor of Science

Rice University

Electrical Engineering, 3.9/4

• Cum Laude
• Distinction in Research and Creative Work

Research Assistant

Stanford Vision Lab

Reinforcement Learning, Imitation Learning, Robotics, Sim2Real

• Built robotics perception and learning infrastructure foundational to the Stanford Arm Farm project, enabling large-scale manipulation research
• Deployed a virtual reality interface to scale task demonstration data collection, enhancing training data diversity and volume for imitation learning
• Developed a MuJoCo-based virtual robot control environment with multi-modal interfaces (HTC Vive, keyboard, mouse) for scalable, crowd-sourced task demonstration collection
• Designed a robot simulation environment in BulletPhysics to support efficient policy training and prototyping for manipulation tasks
• Researched sim-to-real transfer in robotic manipulation using policy distillation and sample-efficient reinforcement learning to improve real-world deployment

Research Assistant

Rice Efficient Computing Group

GoogleNet, CNN, Caffe

• Developed a simulation framework for energy-efficient image classification using the Caffe deep learning library
• Augmented GoogLeNet architecture with custom noise and energy cost layers to evaluate noise-energy trade-offs during inference
• Trained ImageNet models on the modified Caffe framework, optimizing for accuracy under energy and noise constraints
• Designed on-chip memory layout using Cadence tools to support energy-efficient deep learning workloads

Research Assistant

Rice Scalable Health Initiative

Pattern Detection, Motion Recognition, Digital Health

• Developed a wearable stride and motion detection system capable of real-time activity recognition and feedback generation
• Integrated accelerometers, pressure sensors, and storage modules on the MSP430F6726 microcontroller for wearable data acquisition
• Implemented real-time Bluetooth data streaming and digital signal processing in MATLAB, including sensor feature extraction and motion pattern recognition
• Designed, validated, and iteratively refined motion recognition algorithms based on real-world sensor data

RoboTurk: A Crowdsourcing Platform for Robotic Skill Learning through Imitation

CoRL 2018

A convenient remote human demonstration platform for robot task data collection

Neural Task Programming: Learning to Generalize Across Hierarchical Tasks

ICRA 2018

A novel robot learning framework for few-shot learning from demonstration

RedEye: Analog ConvNet Image Sensor Architecture for Continuous Mobile Vision

ISCA 2016

Shifting early vision processing into the analog domain for continuous mobile vision

Robotic Multi-Item Type Palletizing & Depalletizing

US_10,549,928_V1

Robotic Handling of Soft Products in Non-Rigid Packaging

No. 16/797,35

Robot Carriage Tray Table

No. 17/960,544

Multiple Robot Simultaneous and Synchronous Pick and Place

No. 18/393,504

Emulated Robot Peripherals

No. 63/435,176

Team of Excellence

XMotors.ai

Annual award for the outstanding XNGP team

IEEE Eta Kappa Nu Rice Chapter

IEEE-HKN

The honor society of IEEE dedicated to recognizing individual excellence

Louis J. Walsh Scholarship in Engineering

Rice University

Scholarship for deserving students at Rice University

President's Honor Roll

Rice University

Recognition of undergraduate outstanding academic achievement

ELEC Fitness Model Award

Rice University Electrical Engineering Department

Distinction in Research and Creative Work

Rice University

Awarded to the top ten percentile of graduating seniors

Bill Wilson Best Senior Design Award

Rice University

Rice University Affiliates Day student award