Asim H. Gazi

I am a postdoctoral fellow in computer science and statistics at Harvard University, advised by Susan A. Murphy. I completed my Ph.D. in electrical engineering in 2023 from the Georgia Institute of Technology, advised by Omer T. Inan and Christopher J. Rozell.

My research is on sensor-informed closed-loop control systems that can learn to personalize mobile health (mHealth) support to changes in sensor data (e.g., from your smartwatch). These precision health systems enable just-in-time support that adapts to the dynamics of daily life – support that is unaddressed by current healthcare delivery (e.g., therapy for only 1 of the 168 hours per week). My contributions thus far have centered around dynamic modeling, estimation and control algorithms, and data science that account for uncertainties in sensor-informed decision making - uncertainties that are especially prevalent in mHealth (e.g., frequent corruptions of sensor data).

I am currently funded by a NIH K99/R00 Pathway to Independence Award from the National Institute of Biomedical Imaging and Bioengineering, which will fund the remainder of my postdoc and up to $747k over my first three years as an assistant professor at whichever institution I join. Previously, I was supported by Schmidt Sciences, in partnership with the Rhodes Trust, as one of 32 Schmidt Science Fellows selected from around the world in 2023. My PhD was funded by a National Science Foundation Graduate Research Fellowship.

Outside of research, I am a Pathfinder for Boston Partners in Education and the founder and owner of Internet of Tutors LLC.

Research Summary

To help prevent and treat chronic diseases, I work on the next generation of sensor-informed, intelligent mobile health (mHealth) systems. These closed-loop control systems provide personalized healthcare interventions during everyday life, leveraging wearable sensors and other ubiquitous sensor technologies for real-time feedback. My research addresses the need to provide timely support that adapts to the dynamics of daily life – support that is left unaddressed by current healthcare delivery that relies on infrequent, synchronous support (e.g., weekly psychotherapy for less than 1% of one’s life).

The premise of my work is that precision medicine should not just entail tailoring interventions to a person’s traits - an individual’s genetics or characteristics that are (effectively) constant over time. MHealth interventions should also be tailored to a person’s states - biobehavioral context variables such as physiological stress that change over time. A remarkable and challenging aspect of health care delivery is that the same intervention can produce vastly different outcomes even when delivered to the same individual - just because the individual is in a different biobehavioral state.

As an engineer with interdisciplinary research training, my use-inspired research focuses on the algorithms, computational models, and data analytics that enable closed-loop control for intelligent mHealth systems. I work specifically in the area of sensor-informed decision-making to help adaptively tailor mHealth support to data we can passively sense on an individual’s biobehavioral state.

My core engineering contributions have centered around methods to manage uncertainty when inferring changes in biobehavioral state or personalizing sensor-driven interventions. Sources of uncertainty include (1) intermittent corruptions of wearable sensor data, where signal quality assessment is paramount; (2) fallible predictions of biobehavioral states, where uncertainty-informed decision making is vital; and (3) the lack of quantitative mechanistic models to inform decision making for nascent mHealth interventions such as non-invasive vagus nerve stimulation, where hybrid modeling and digital twins are important to inform control design and system deployment.

I have also been fortunate to collaborate with industry partners and clinician scientists on translational efforts bridging research and healthcare delivery. For example, collaborative research with the Emory School of Medicine formed the scientific basis for US Food and Drug Administration Breakthrough Designation in 2022 of non-invasive vagus nerve stimulation for patients with posttraumatic stress disorder .