About

Yasser Khan is an Assistant Professor of Electrical and Computer Engineering and of Biomedical Engineering at the University of Southern California, where he joined in 2022. He holds the Andrew and Erna Viterbi Early Career Chair. Dr. Khan earned his B.S. in Electrical Engineering from the University of Texas at Dallas, his M.S. from KAUST, and completed his Ph.D. in Electrical Engineering and Computer Sciences at the University of California, Berkeley, in 2018. He also conducted postdoctoral research in the Department of Chemical Engineering at Stanford University. His research centers on additive manufacturing and hardware-enabled AI, with a focus on developing next-generation wearables, implantables, and ingestibles for precision health and psychiatry. Dr. Khan has received numerous honors, including the 2024 Packard Fellowship, the 2025 Air Force Office of Scientific Research Young Investigator Award, the 2025 IEEE Sensors Council Early Career Technical Achievement Award, the 2025 USC Viterbi Junior Faculty Research Award, and the 2023 Google Research Scholar Award. His work has been published in prominent journals such as Nature, Science, Cell Reports, Advanced Materials, and PNAS, and has been featured by major outlets including BBC News, The Wall Street Journal, Newsweek, and NSF News.

Research topics

• Computer Science
• Materials science
• Artificial Intelligence
• Nanotechnology
• Embedded system
• Engineering
• Physical medicine and rehabilitation
• Mechanical engineering
• Chemistry
• Biology
• Knowledge management
• Human–computer interaction
• World Wide Web
• Marketing
• Computer vision
• Physics
• Speech recognition
• Electrical engineering
• Business
• Electronic engineering
• Mathematics
• Neuroscience
• Psychology

Selected publications

• Self-severing circuits facilitate passage of ingestible electronic sensor-guided therapeutics

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-03-30

  articleOpen access

  Ingestible electronics enable the tracking and treatment of gastrointestinal and systemic diseases. However, bulky batteries and circuit boards require large capsules that can result in bowel obstruction, a medical emergency. Here, we engineered a 9 × 26 mm electronic pill capable of triggered severing into tiny pieces with sizes clinically proven to reduce obstruction risk. Our capsule enables multicomponent circuit boards to connect with separately encapsulated powering elements via conductive, interlocking connections. Heat induced softening of polyethylene glycol/polycaprolactone channels activates a spring to separate encapsulated components into inert 9 × 15 mm segments, facilitating intestinal passage. Separation triggers included closed-loop sensors and time-delay circuits. In vivo swine studies demonstrate the ability of our capsules to sense luminal oxygen changes via an optoelectronic sensor, locally trigger upadacitinib delivery, and facilitate safe excretion.

  PublisherOA PDFDOI

• Investigating a User-Friendly Machine Learning Approach to Predict the Gas Hydrate Formation Temperature

  Journal of the Geological Society of India · 2026-04-01

  article

  ABSTRACT The petroleum industry faces a serious problem of gas hydrate formation in pipelines and process equipment, particularly in low-temperature marine environments. It is necessary to understand the chemical thermodynamics of gas hydrate formation so that it can be conveniently avoided. Although research has been conducted to develop correlations or to use Artificial Intelligence (AI) to determine hydrate temperature, to the authors’ knowledge, none have compared Machine Learning (ML) models’ performance in hydrate formation temperature prediction, especially Web User Interface (WUI) development based on the best-performing model. To achieve precision, this work compares the performance of four ML models, such as Artificial Neural Networks (ANNs), Decision Tree Regressor (DTR), Support Vector Regressor (SVR) and Random Forest Regressor (RFR) in terms of hydrate formation temperature prediction using operating pressure and specific gravity as features. Python was employed for this work, as it supports open-source libraries such as Keras with TensorFlow and scikit-learn, among others. Results showed that the coefficient of determination (R2), Root Mean Square Error (RMSE), and a20-index on the overall dataset are (0.9994, 0.3120, 1), (0.9992, 0.3716, 1), (0.9991, 0.3880, 1) and (0.9910, 1.2545, 0.9972) for DTR, ANNs, RFR, and SVR, respectively. Although all models delivered strong results, they differ sharply in the time required to train. ANNs required 974.9498 s, RFR needed 15.9753 s, and SVR took 294.4242 s. In contrast, DTR completed the task in only 0.2727 s. Based on performance and computational efficiency, the models rank as follows: DTR>RFR>SVR>ANNs. Eventually, Web User Interface (WUI) was developed based on the bestperforming model (DTR). The optimal activation function for the ANN is tanh, while the Support Vector Regressor model performs best with the Radial Basis Function (RBF) kernel. We are optimistic that this research will open novel avenues in natural gas engineering. It is recommended that the models’ lower and upper bounds be broadened by training the model on additional experimental data across different operating conditions.

  PublisherDOI

• Complementary visual localization and tactile mapping approach for robotic perception of millimeter-sized objects with irregular surfaces

  Microsystems & Nanoengineering · 2026-03-11

  articleOpen access

  Humanoid robots and human-machine interaction technologies are essential for perceiving and manipulating millimeter-scale objects with irregular surfaces in extreme environments, such as outer space, radioactive zones, and hazardous sites with explosive ordnance, where human access is restricted. A vision-based perception approach provides spatial and positional information about objects but relying solely on it for robot manipulation poses challenges due to limitations in detectable object size, as well as sensitivity to external factors such as focusing issues, occlusion, and lighting conditions. In contrast, tactile perception offers valuable information about aspects that are difficult to discern visually, including an object's shape, surface characteristics, and the forces involved during contact. This study presents a complementary visual localization and tactile mapping framework that allows robots to effectively perceive small objects with irregular surfaces in visually restricted environments. The proposed method draws inspiration from the sequential vision-tactile sensory processing observed in humans when handling small objects with irregular surfaces. It employs an RGB-Depth camera for visual perception and a soft pressure sensor array, made using inkjet printing, for tactile perception. We demonstrate the feasibility of implementing a sensory substitution to detect the size and location of objects through visual perception, as well as identify object surfaces and reconstruct their three-dimensional profiles using tactile scanning, particularly in environments where visual information is limited. This study provides a technological foundation for enhancing the autonomy and adaptability of humanoid robots in unpredictable and unstructured environments, particularly to support precise robot manipulation in such conditions.

  PublisherOA PDFDOI

• OxyJet CPAP: an electricity-free low-cost emergency respiratory support device

  BioMedical Engineering OnLine · 2026-03-04

  articleOpen access

  Respiratory support devices in resource-limited settings should be inexpensive, portable, effective, and easy to use. Non-invasive respiratory support devices can reduce expensive ICU admissions, but these facilities are severely lacking in low-resource settings. Here, we describe the design and validation of a low-cost, portable, electricity-free, and 3D-printed continuous positive airway pressure (CPAP) device named ‘OxyJet’ that can provide non-invasive respiratory support outside the ICU. The OxyJet is uniquely built using off-the-shelf components and 3D printing technology, making it both inexpensive and easy to produce. OxyJet costs less than 10% of the price of a similar CPAP system. Inspired by the fundamental mechanics of gas ejectors and harnessing the potential energy of a high-pressure oxygen jet, the device can deliver a high flow of oxygenated air up to 65 L/min (approx.), a positive end-expiratory pressure (PEEP) within 5–15 cmH2O, and a fraction of inspired oxygen (FiO2) of up to 100%. The device was bench-tested following UK-MHRA RMCPAP guidelines and tested on healthy volunteers (n = 5) and hypoxemic patients (n = 5). A comparative pilot study involving 23 hypoxemic adult patients conducted in Dhaka, Bangladesh, showed a significant improvement (p < 0.05) in the peripheral oxygen saturation (SpO2) of patients following the administration of OxyJet CPAP. The mean SpO2 increase was 12.0% (95% CI 10.8–13.2) with OxyJet versus 11.5% (95% CI 9.3–13.8) for standard CPAP (p = 0.695). The findings indicate the device's feasibility and short-term physiological effects comparable to those of standard CPAP systems. Further studies are required to confirm its clinical efficacy and broader utility in resource-limited settings. Our findings suggest that OxyJet CPAP has the potential to serve as an emergency respiratory support device outside the ICU, strengthening health systems in resource-limited settings. OxyJet’s performance was first assessed through benchtop testing following the UK-MHRA RMCPAP protocol. This was followed by preliminary human testing in healthy volunteers and hypoxemic patients to evaluate safety and usability. A pilot feasibility study involving hypoxemic adult patients in Dhaka, Bangladesh, was then conducted to compare the device’s physiological effects with those of standard CPAP therapy.

  PublisherDOI

• Single-chip End-to-End Ingestible Electronics for Gut Neurotransmitter Sensing

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-03-31

  articleOpen accessSenior authorCorresponding

  Abstract Neurotransmitters in the gut play a vital role in human health and neuroscience, and their real-time monitoring is essential for understanding underlying physiological mechanisms. However, bioelectronic systems capable of measuring neurotransmitters in vivo at the anatomical site of interest remain underdeveloped and largely depend on bulky, off-the-shelf electronic components, thereby constraining the development of systems that are both practical and minimally invasive. Here, we report a miniature ingestible pill that is capable of real-time in vivo sensing of two key neurotransmitters: serotonin (5-HT) and dopamine (DA). The system incorporates a fully printed three-electrode-based electrochemical sensor for neurotransmitter sensing and a custom application-specific integrated circuit (ASIC) that integrates all major functional blocks on a single chip, enabling a platform for fully wireless monitoring of gut neurotransmitters. The pill, measuring 5.8 mm in diameter and 19 mm in length, supports multiple electrochemical sensing techniques, including amperometry and voltammetry, with only 42 μ A of average current consumption. We demonstrate the ingestible platform through in vivo studies in rat animal models, enabling real-time monitoring of gut neurotransmitters.

  PublisherOA PDFDOI

• Improved dynamic MRI of the wrist and heart at 0.55 T enabled by rapid 3D printed flexible coils

  Nature Communications · 2026-04-20

  articleOpen accessSenior author

  Dynamic MRI at low field is limited by reduced signal-to-noise ratio (SNR) and additional losses imposed by accelerated acquisitions required for high temporal resolution. Flexible, anatomically conformal coils can recover SNR by maintaining proximity during motion, but lower conductor conductivity introduces resistive losses, requiring careful material optimization. We demonstrate high-quality dynamic imaging at 0.55 T using coils fabricated via two rapid, low-cost digital methods: direct-ink-write silver printing and screen printing of copper-doped EGaIn. Four-element wrist arrays were fabricated in 8 minutes per element at approximately $30 in consumable materials. In three subjects, both arrays achieved up to 4 × higher contrast and 5 × greater sharpness than a commercial coil during dynamic wrist imaging. To demonstrate generalizability, a seven-element silver-ink cardiac array produced cine images comparable to a commercial coil. These results establish digitally fabricated flexible coils as a scalable and accessible solution for dynamic musculoskeletal and cardiac MRI at low field.

  PublisherDOI

• An Empirical Model of Ion‐Selective Organic Electrochemical Transistors (Adv. Mater. Technol. 2/2026)

  Advanced Materials Technologies · 2026-01-01

  articleOpen accessSenior author

  Ion-Selective Organic Electrochemical Transistors In their Research Article (10.1002/admt.202501217) Mainul Hossain, Yasser Khan, and co-workers present an empirical model for ion-selective organic electrochemical transistors (IS-OECTs) that predicts sensitivity and selectivity across varying ion concentrations in biofluids such as sweat. The model captures steady-state behavior in the presence of both target and interfering ions and accurately reproduces concentration-dependent responses, demonstrating strong reliability for advanced electrochemical sensing. Art by Bayazid Bulbul.

  PublisherOA PDFDOI

• Wearable organic-electrochemical-transistor-based lithium sensor for precision mental health

  Device · 2025-07-18 · 2 citations

  articleOpen accessSenior author
  PublisherOA PDFDOI

• A Wearable Sweat Rate Sensor With Adaptive Sweat Ion Concentration Calibration

  IEEE Sensors Letters · 2025-06-12 · 3 citations

  articleSenior author

  Wearable sensors for continuous physiological biomarker monitoring offer a non-invasive and personalized approach to healthcare. Sweat, as a readily accessible biofluid, serves as a valuable medium for biomarker detection. However, accurate biomarker analysis relies on precise sweat rate measurements, as sweat concentration and secretion rates are interdependent. To address this, it is critical to calibrate sweat ion concentration to enhance the accuracy of sweat rate monitoring. Moreover, the distribution of sweat glands varies across different body regions, necessitating application-specific design modifications for wearable sweat rate sensors. Effective sensor implementation requires adaptability in design while ensuring compatibility with scalable fabrication techniques. In this work, we present a digital fabrication method that allows for easy customization and large-area manufacturability. A key advancement in our approach is the integration of secondary capacitive sensing to continuously monitor sweat ion concentration, which is incorporated into real-time sweat rate calculations. This calibration technique enhances measurement precision by accounting for fluctuations in sweat composition. Additionally, we developed customized readout electronics and a mobile application for real-time sweat rate visualization, enhancing user accessibility and convenience. By incorporating sweat ion concentration calibration through capacitive sensing, our sensor system significantly improves the accuracy and reliability of sweat rate measurements, advancing real-time perspiration monitoring for personalized health applications.

  PublisherDOI

• A Flexible MRI Array based on Direct-3D-Write Technology at 0.55T

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  articleSenior author

  Motivation: MR receiver arrays are typically made from rigid copper which is slow to manufacture and cannot fully conform to imaging anatomy or adapt smoothly during movement. Goal(s): To demonstrate a flexible 4-element array that can conform to target anatomy, can perform well with parallel imaging, and is fast and inexpensive to manufacture Approach: We utilize highly conductive silver-based ink and a fast direct-3D-write method that uses easy to modify components. Results: The flexible array provided good targeted coverage, adequate array decoupling (&lt;7% coupling between elements), and satisfactory parallel imaging performance (average g-factors: 1.1 for rate-2 and 1.4 for rate-3) Impact: Flexible MRI receiver arrays, created using 3D-write technology, can be made to conform to the target imaging anatomy while also offering scalability, low cost, and quick manufacturing. They are conductive enough to provide adequate SNR for effective parallel imaging.

  PublisherDOI

Frequent coauthors

• Ana Claudia Arias

  University of California, Berkeley

  27 shared

• Jonathan Ting

  University of California, Berkeley

  11 shared

• Natasha A. D. Yamamoto

  University of California, Berkeley

  9 shared

• Md. Farhad Hassan

  University of Southern California

  8 shared

• Donggeon Han

  7 shared

• Felippe J. Pavinatto

  GE Global Research (United States)

  7 shared

• Luca Benini

  ETH Zurich

  7 shared

• Zhenan Bao

  7 shared

Education

• PhD, Electrical Engineering and Computer Sciences

  University of California Berkeley

  2018

Awards & honors

• 2024 Packard Fellowship
• 2025 Air Force Office of Scientific Research Young Investiga…
• 2025 IEEE Sensors Council Early Career Technical Achievement…
• 2025 USC Viterbi Junior Faculty Research Award
• 2023 Google Research Scholar Award