• Led end-to-end development of a speech deepfake detection system leveraging wav2vec audio encoder and a Mamba state space classifier model; achieved 6.5% Equal Error Rate on an unseen dataset, reaching state-of-the-art performance and enabling financial institutions to proactively detect fraudulent calls in customer support operations.

• Architected and deployed a real-time speech emotion recognition system by modifying wav2vec and applying post-training regularization (batch-level variance maximization, dimension-level covariance minimization), improving six-class accuracy to 88%. Enabled quantification of customer emotional states, enhancing agent empathy and adaptive support to boost customer satisfaction and team professionalism.

• Built Digital signal processing data augmentation pipelines to simulate diverse recording conditions, ran W&B hyperparameter sweeps, and deployed the system on AWS/GCP for scalable, robust performance.

• Analyzed dataset domain shift through signal processing techniques and comparative experiments, generating insights into cross-dataset robustness and failure modes for fraud detection.

• Designed and implemented a real-time control and perception pipeline for the Unitree quadruped robot, integrating 3D point cloud processing and height map generation to support terrain-aware locomotion and future autonomous navigation.

• Developed transformer-based motion and exploration models for metric-free topological mapping, achieving a 95.1% exploration rate via imitation learning in feature space

• Utilizing few-shot prompting to empower a QA agent to autonomously generate code and query scene-level data from images for Embodied QA tasks.

Reinforcement Learning and Games Lab

• Applied self-supervised consistency techniques to enhance and stabilize hidden state representation learning for the DeepMind MuZero RL model, enabling the agent to achieve superhuman performance in Atari games with fewer interactions.

• Implemented a TD learning RL model with tree search for the stochastic dynamic environment-based computer game “2048.”

• Developed a visualization tool in Java that facilitated better analysis of the RL models’ performance across different board games.

• Working under the supervision of Research Fellow WU, TI-RONG