Skip to main content
Amanda Randles

Amanda Randles

Computer scientist

Aug 24, 2026

18:40
How did you come up with that question at that time? And give us a little context for that because I know you told us now you're really involved in hemodynamic modeling using the principles of fluid dynamics in your work.
18:56
Yeah.
18:58
So I think that's where...
19:00
So I started grad school in about 2009, 2008.
19:04
So I started the Blood Flow Project about 2009.
19:07
So I joined...
19:09
Basically, it's been a while and things have developed a little bit.

11 MINS LATER

30:14
Like Tell us a little bit about what happens at the petabytes per time set scale.
3:28
Based on your work, where are digital and virtual twins already providing the most concrete value in predicting device performance across patient populations?
3:40
So I think right now the clearest value is really where, you know, it, it... places where the, the device performance depends strongly on that patient's anatomy, where it's the geometry of their coronary arteries or maybe their valve geometry or whatever is going to really affect how that device is going to play out.
3:56
So typically it's gonna be things like cardiovascular devices, vascular interventions.
4:01
We see it a lot for stents, for valves, for flow diverters, ventricular assist devices, and a lot of these, like, pressure and flow-based diagnostics.
4:09
So one of the, the key examples we see before you need the device is to determine if you actually need the device.
4:14
We're seeing a lot for fractional flow reserve.
4:16
So you have companies like HeartFlow and CathWorks similar, um, that are already getting FDA approval, that are already out there being used, really helping you.

8 MINS LATER

12:27
it.Well, looking ahead, how do you see advances in AI, high-performance computing, and data availability accelerating simulation-first prototyping, and could this fundamentally reshape the economics and timelines of medical device innovation over the next decade?
Margaret HarrisCORRESPONDENT
7:08
So how does it work to sort of link what, what's the timescales on what you're linking the information you're gathering from wearables, you're gathering from scans, and how are you up- using it to update the digital twin so you get information about the, the actual person?
7:21
Yeah, no, that's a great question.
7:22
So if-- The way, w- the way we're doing it is that we really see it as, you know, a blending of data from things like the wearable sensors with the physics-based modeling.
7:29
What we're trying to set up right now is we're-- we've developed this method, we call it longitudinal hemodynamic mapping framework, and it allows you to really kind of take a parallel on time approach to simulating the hemodynamics.
7:40
And it allowing...
7:41
Like, basically, we're, we're, we're getting at, you know, a finite number of hemodynamic units that we can identify, that if you can piece those back together and, and, you know, figure out, it's almost like a lookup table of, you know, you figure out what these hemodynamic units are, you use your wearable data, and you figure out which hemodynamic unit you should be pulling from.
7:56
What we need to do then is simulate all of those hemodynamic units when we first get the medical imaging, so we're trying to minimize the amount of time it takes, you know, the number of hemodynamic units and how efficiently we can calculate each of those.
Margaret HarrisCORRESPONDENT
12:42
What's that look like for you?

We value your privacy

We use cookies to understand how you use our platform and to improve your experience. Click “Accept All” to consent, or “Decline non-essential” to opt out of non-essential cookies. Read our Privacy Policy.