Khan Lab at USC | Los Angeles, CA | Jun. 2023 - Aug. 2025

Khan Lab at USC Research Experience

Built the software and embedded-systems layer for wearable cognitive-state research, spanning experiment interfaces, multimodal physiological sensing, real-time dashboards, and machine-learning analysis.

Khan Lab Posters

Research role

Individual Researcher under Dr. Khan, Ming Hsieh Department of Electrical and Computer Engineering

My work at Khan Lab sat between biomedical sensing and production-minded engineering. I helped turn lab protocols into repeatable software workflows: collecting synchronized physiological signals, validating custom hardware, building dashboards for live inspection and recording, and preparing structured datasets for stress and affect classification.

Research systems

What I built

Cognitive stress evaluation and prediction platform

I worked on a Python-based platform for running cognitive-state experiments and converting wearable biosensor streams into machine-learning-ready data. The system supported relaxation, cold pressor, Stroop, arithmetic, and video stimulus tasks while logging participant metadata, timestamps, and task durations into structured files for analysis.

Integrated EDA/GSR and PPG-derived signals, including conductance, heart rate, and heart-rate variability features.
Built experiment interfaces for repeatable cognitive tasks and event logging so data could be aligned with protocol stages.
Processed recordings in Jupyter with NumPy, Pandas, and SciPy for merging, trimming, feature extraction, and dataset cleanup.
Evaluated classical ML baselines including random forest, KNN, and SVM; poster results reported 95 percent accuracy for random forest and KNN, and 87 percent for SVM.

Multimodal physiological acquisition firmware and dashboards

I also contributed to the data acquisition stack behind multimodal wearable sensing. The platform combined microcontroller firmware, sensor communication, Python tooling, and a WebBLE dashboard so lab members could inspect, record, and validate physiological streams in both wired and wireless workflows.

Supported firmware development around Arduino Nicla Sense ME with ADS1299 EEG/ECG acquisition over SPI.
Worked with I2C-connected AD5593R EDA and MAX30102 PPG sensing paths for multimodal signal capture.
Built and validated dashboard workflows for serial visualization, recording, wireless DAC control, and EEG sensor checks.
Helped design a WebBLE dashboard for wireless monitoring and logging of PPG, EDA, and movement data.

Outcomes

Why it matters

Turned exploratory neuroengineering experiments into repeatable software workflows with structured logs and analyzable datasets.
Connected embedded sensing, live visualization, and offline ML analysis into one research pipeline.
Added a biomedical-signal thread to my AI systems background: noisy sensor data, constrained hardware, data quality, and model evaluation.

Stack

Tools and methods

PythonC++JupyterNumPyPandasSciPyscikit-learnRandom ForestKNNSVMEEGPPGEDA/GSRSPII2CWebBLE

Posters

Research artifacts

Khan Lab neuroergonomics and wearable sensing poster