Dr. Ayushi Sinha, PhD in Computer Science | Senior Scientist at Philips

In this episode, we discuss Ayushi’s experience pursuing her doctorate in Computer Science at Johns Hopkins, her advice for finding an advisor and an academic environment that supports your unique interests and needs, and the different ways she applies her computer science skills in the medical field working with x-ray imaging systems as a Senior Scientist at Philips.

Hosted by Megan Benay

To connect with Ayushi and to learn more about her story, visit her page on the PHutures #100AlumniVoices Project website.

Transcript

Megan Benay

Hi. I'm co-host Megan Benay, and this is the 100 Alumni Voices Podcast, stories that inspire, where we explore the personal and professional journeys of a diverse group of 100 doctoral alumni from Johns Hopkins University. Today we're joined by Doctor Ayushi Sinha, PhD in computer science and is currently a senior scientist at Phillips Research. Doctor Sinha, welcome.

Ayushi Sinha

Thank you so much, Megan. Great to talk to you.

Megan Benay

So, you are currently working at Phillips Research and when we're talking about Phillips, we're talking about the company that the layperson such as myself might know as the the group that makes products like our electric toothbrushes and TV's. Is that correct?

Ayushi Sinha

That’s correct. Yeah, Phillips has been or in the past at least kind of has been like the Google of electronics where it used to be involved in all sorts of domestic appliances and oral healthcare or personal healthcare. Like they also make razors. Yeah, they also used to make light bulbs. Actually, Phillips started off as a light bulb company. But in the past few years, I want to say about 10 years they have been kind of divesting their sort of non-healthcare businesses to other companies, so they sold their lightning or lighting arm and then recently sold their domestic appliances arm and now are completely focused on healthcare. So, although they are known for their toothbrushes or light bulbs, they actually also make MRI machines and CT machines and X-ray ultrasound. And so that's where my work is involved.

Megan Benay

OK, so where is your work involved if it's not in making electric toothbrushes?

Ayushi Sinha

I work mostly with X-ray imaging systems and the area that I'm I'm involved in is is what we call image guided therapy. So basically, it's minimally invasive interventional care where the patient is treated from a very small incision. So, sort of different from the past where if you were treating a cardiac ailment or something in the brain you would open up the patient and do an open heart surgery or you would open up the patient’s head and sort of directly treat the impacted area, whereas now just because of the complications involved and recovery period involved in opening up a patient, most treatments are kind of shifting towards a minimally invasive approach where you make a small incision either in the groin or in the arm, and you insert wires essentially that are sort of the devices that I use to sort of traverse the vasculature of the patient to get to the site that is going to be treated. And so, as you can imagine, the issue with this is that you haven't opened up the patient, so you can't look inside. And so, we use X-ray to look inside and see where the devices are and if we are placing devices in in the right location and basically providing treatment at the right site.

Megan Benay

OK, so forgive me because everything you just said sounds like you maybe went to medical school, but if my if my research has been correct you studied computer science at Johns Hopkins. So, are you? How does computer science play a role in what you're doing today?

Ayushi Sinha

Yeah, that's a great question. And and I think also great knowledge to bring to a lot of people that are good at writing code and want to pursue computer science that you can sort of apply your skills in many different fields and I ended up sort of in the medical field and I think obviously having gone to Johns Hopkins, that's probably a great influence on that. But the computer science department at Johns Hopkins collaborates very closely with the Hopkins Hospital. So basically, we work on projects where we bring things like image processing together with treatment delivery. So, in my current position, I look at 3D or 2D images. So, these X-ray systems can acquire both 3D images, thus they're spinning around the patient or 2D images from sort of one source location. And basically, the idea is that you want to take away the tedious steps involved in the procedure where the physician has to look at these complex 3D images and make certain decisions and and and instead use image processing to either kind of lead them in the right direction or give cues as to how best to treat the patient. So that's where my computer science skills come in in analyzing the images and making recommendations to the users.

Megan Benay

And so, does that start to automate some of the decision making? Is that right?

Ayushi Sinha

Correct, yeah. So, and there are also a lot of steps in the procedure that aren't necessarily sort of direct decision making, in the sense that sort of similar to how our phones are becoming more and more complex and more capabilities available to the point that I probably use only 10% of the capabilities on my phone, right? I'm not using everything that it provides. And that's a similar trend that's been seen in the operating room where imaging systems and all of the devices that are present in the room are becoming more and more complex and provide more and more capabilities. But physicians can't keep up a) because it's hard to kind of remember everything that it offers and b) they need to operate on the patient and not on the imaging system. So, they can't spend their time checking all of the settings and checking all of the available features. So, the idea is that we kind of understand the context of the procedure and make particular suggestions on what settings to use to get the best image, for instance. So, not direct decision making, but if they get the best image then the assumption is that they should make the best decision based off of that image.

Megan Benay

Seems like it maybe walks the line or share shares a boundary, perhaps with artificial intelligence and one of the one of the pieces of artificial intelligence I'm sure you're aware of and that we're all talking a lot about in academia right now is ChatGPT? And I'm curious if you have any any computer science insight into that or any any knowledge opinions where you think we're going with artificial intelligence in academia?

Ayushi Sinha

Yeah, both in academics and in industry there's a lot of emphasis, of course, on using AI or using machine learning and deep learning to do things that we weren't able to do in the past. So, we use a lot of deep learning algorithms, for instance, also to analyze the 3D images that we acquired or the 2D images that we acquired just as examples. And I think some of the main concerns or things to keep in mind similar to ChatGPT and other, you know deep learning techniques that are used in language processing also apply to image processing, which is that we need to keep in mind that the data that we use to train our models are very sort of what kind of data that is is very important to keep track of and make known. Because there are a lot of biases that exist in images as as well as in language. There are a lot of assumptions that may be made given a particular demographics that the images are coming from that don't necessarily apply to other demographics, for instance. Different hospitals follow different procedures. They have different protocols. So, I think the most important thing is to kind of understand where the data is coming from and keep track of the biases, because if we don't, then we might be able to show certain results in our in our algorithms based on the data that we acquired, which a lot of the data that we get at Phillips are from some of the sort of best sites in the world. So, we're basically getting images that are acquired by some of the best physicians in the world. So, we need to keep that in mind when we translate those algorithms to sites that are just up and coming, and make sure that any assumptions that we made at these sort of you know top sites are sort of analyzed in the context of newer sites or sites where different protocols might be followed.

Megan Benay

That's an interesting it's interesting you bring that up around bias, biases and data. And I know that. Well my my background in data in schools, we work a lot around how are we how are we using this data to break down barriers and and trying to dismantle, you know, implicit and explicit bias. I'm curious if in your career trajectory or with data, have there been any times where you've really been confronted with a problem that you're like, hey, this data is woefully biased. We shouldn’t be using this. Or where you've tried to combat it and how that went for you?

Ayushi Sinha

That's a good question. I think some of the images that we use may be at least from what I can tell right now, may be less susceptible to bias just because if you X-ray, you know a bone in a man or a woman, the bone, the bone structures do vary depending on genders, and specifically particular types of bones like pelvic structures, are different, but if you're looking at bone in the arm you can't really tell what's different or that it came from a man or a woman. I think the modality that we use kind of almost protects us from some of the biases, but that isn't to say that we don't have any bias in the data right. There may be biases that we don't even realize exist, so it's a little bit difficult to say what specific bias we have encountered and overcome, but the sort of thing that I've noticed that has been kind of positive in my experience is that everybody is talking about bias. So, there are people on our teams that are specifically served tasked with finding sites across the world, so we have a site that we work with in North America and a site that we work with in South America. There's a site that we're trying to find in Europe, a site in Asia. So, we're we're trying to sort of diversify the sources of our data. So at least I think there's already a proactive sort of attempt to make sure that we are combating, at least you know, demographic based sources of bias.

Megan Benay

And Speaking of Speaking of biases, you are a woman in data and computer science, and that in and of itself is, you know, you are a minority in terms of our gender or I assume we share we share our gender affiliation and that is something to get to where you are today, to be a woman in this field, and I don't know if you are comfortable talking at all about your any other identities that you may have that may where you have also experienced any kind of marginalization, but how did you how did you combat that? I mean, I know for me I sit in spaces that are pretty much all men. Data, even in education, education, which is full of women, data in education is a lot of men. And I sit in the room with a lot of men and I can imagine for you, it is the same if not amplified. So, how have you handled that?

Ayushi Sinha

Yeah, the the answer to that question, I feel like is very, very complex. And I feel like I've noticed.

Megan Benay

Like just a lifetime to unpack.

Ayushi Sinha

Yeah, the topic itself, right, is so complex. And not only do different people sort of experience this in different ways, I feel like at different points in my life I've also experienced it sort of differently. Growing up, I feel like I never really thought about it, even though I mean these sort of biases and assumptions that you know men are better in engineering or other spaces kind of persist everywhere, but I I feel like either I was naive or or something but so growing up I never really noticed it. I grew up in India and I think in India there's a sort of general assumption that if you want to succeed in life, you either become an engineer or a doctor. And so, I kind of heard that.

Megan Benay

If your woman? If you're a woman or a man? Just both?

Ayushi Sinha

Just just in general. If people were implying that this was gendered, I missed it. But at least you know within my family there was a lot of sort of push to you know, be good at math and like science, so I kind of just went in that trajectory and, you know, somewhere down the line realized that I liked writing code. So, I kind of just followed that. And never really felt like I was the only one following that path. Until I think, I think at every sort of stage of higher education, I maybe started to realize that mSore and more. So, when I went to College in a lot of my computer science classes, I was the only or one of two women. And so that's when I really started to notice that, you know, there's this imbalance. And then in Graduate School, there were more women, but I think there were a lot of also assumptions on experience, or you know confidence and the environment may not have been as sort of set up for for somen, as as it was for men. And I don't know if I can necessarily give particular examples, but I definitely did have experiences where there were certain things assumed of of me and of my experience that weren't true, and I felt like I wasn't given the opportunity to learn. And again, whether this was because of my gender or not, I don't know. But there were also people that kind of did help, you know, help me to pick up things that I didn't know. So, when I first joined graduate school I I felt like I was already supposed to know how to do research, whereas I came in assuming that I would learn how to do research and so I felt like I had to kind of figure out who my mentor would be who would really, sort of help me figure this out, and eventually I did find an advisor who really did help me learn how to do research and I feel very confident in my abilities now to do research, but.

Megan Benay

Find your own mentor. Just sorry to interrupt, but that's another I I've read actually a lot about how the there's also a gender difference in how men and women go about seeking out mentorships. You know, people who are going to mentor them or advise them. And and I'm curious, did you did you find a mentor? Were you assigned an advisor?

Ayushi Sinha

I was assigned an advisor, but then I switched advisors, so I did have to kind of look around and figure out who had a project available that I could join. And I I don't know that I necessarily picked an advisor because I could, you know, sense that they would be good at helping, you know, teach someone who had, who had little research experience how to become a good researcher. But I got lucky and in that I ended up with an advisor who sort of handheld when needed, and then let me go and and kind of explore the space that I was starting to get comfortable with. But I can see how that experience can be very intimidating, and to me it was extremely intimidating to the point that I considered, you know, not continuing in the PhD program.

Megan Benay

I imagine that is an experience that many of us have felt. Did you feel like as you were then meeting with with additional folks, did you feel like you needed to sort of like disclose or advocate that you were looking for someone who's going to be able to help you with the research side of things? Like, did you go about saying like, hey, I, you know, I'm asking you, but and you're, you know, you're interviewing me, but I'm also interviewing you, because I'm paying for this program. I'm looking for someone to shepherd me through. So, did you have to sort of like layout also what you were looking for and what you needed?

Ayushi Sinha

Not exactly.

Megan Benay

OK.

Ayushi Sinha

I think just because I feel like it's a very vulnerable position to put yourself in at a competitive program like the computer science program at Hopkins and kind of be able to comfortably say that you also need to learn how to do research. And maybe that assumption is OK now because I feel like a lot of students coming in when I was finishing up and reviewing applicants, everyone seemed to have so much more experience than I did going into grad school, but I think in order to kind of keep sort of the gates open for people coming from, you know, not top research universities. You know, I came from a liberal arts school, but I had degrees in computer science and math, and I know that I was good at these subjects. But I didn't have as much research experience and I feel like keeping the door open for people with that experience is also important because, you know, people with different backgrounds come in with different ideas, and that is sort of the whole point of bringing in people from different backgrounds and allowing that sort of research to occur. So yeah, so again, I don't know. I don't think that any of this was because of my gender. But I I think also how the environment sort of makes men and women feel comfortable kind of makes a difference in how you know, comfortable I felt reaching out to people versus just kind of closing and saying I can't reach out to people, so I'm going to leave. So, I think making environments open to people of different types is helpful.

Megan Benay

So, for our, for our listeners who are maybe like right now in the throws. I remember when when we were assigned our advisors, it was incredibly stressful. People were freaking out about it, and I can imagine it's the same in many programs. Do you have any advice then for folks who have been assigned an advisor and it doesn't feel right or they’re having thoughts like what you're having or what you had about like maybe this isn't right for me. Obviously, you stuck with it and you decided to pursue, you know, seeking out other advisors. Any other advice you have for folks who might be in a similar position as you were in?

Ayushi Sinha

I think talking to as many people as you are comfortable, obviously starting with people that you already work with that you know that can help you. So that's that's what I did. There were a few other professors that I was also kind of working with on different projects, as I was trying to figure out where I wanted to land. So, I talked to some of them both in in terms of whether I could sort of switch to their projects and to their labs, or also who else they would recommend I talk to. And I kind of just follow that trail and you know, talk to a person A they're recommended BCD. And I talked to BCD they recommended others and I kind of just repeated that a few times until I felt comfortable. Or maybe until I felt like there was a project that I could try to contribute to. But I know that that's not easy. You know, professors, especially early on during during your PhD, can be intimidating. And so, I understand by reaching out is not easy, so maybe to kind of turn the recommendation away from students and maybe to programs is to institutes or rotations and programs that allow students to kind of get a feel for how different labs function and where their comfort lies and you know where they have the best match. A lot of departments already do this, and I think it's I think it's good practice. It lets people kind of find their space.

Megan Benay

All right, you heard it here, JHU programming folks. This is the recommendation. And if if you're alumni or your or you’re current students who are who are listening and your program isn't doing this, maybe it's worth banding together and advocating for it. 

Ayushi Sinha

For sure. 

Megan Benay

When you were going, when you were graduating, then did you imagine that you would had you already sort of forged this path into the healthcare field and you knew that you wanted to keep going in that direction or were you considering other things? Like how did your how did your study then your studies at Johns Hopkins, how did that inform your next step?

Ayushi Sinha

So basically, once I sort of switched to a new advisor, their projects were in the medical field. They collaborated very closely with a surgeon at the hospital. And that's really how I started to get into the medical field. So, I really started to even pay attention to the needs that computer science serves within the medical field and became really interested in it. So, I was my PhD project was tackling a very specific type of procedure, which is endoscopic sinus surgeries. So basically, accessing the sinuses through the nose, usually with an endoscope. And then yeah, very interesting. But also, a very not necessarily immediately welcoming space for someone who hadn't really been in the medical field for a while.

Megan Benay

So, it's a, it's a closed club of sticking things ups people’s noses doctors.

Ayushi Sinha

Yes, there is.

Megan Benay

Alright, noted.

Ayushi Sinha

But yeah, in a similar sort of vein, these surgeons need to see what they are operating on, even though they're not opening up the patient. So, they use an endoscope which sends back images from inside the nose, inside the sinuses. And then basically you want to be able to evaluate what you're looking at, and if you're, you know, close to a major artery, then you want to avoid injuring it and things like that. So, that was really my first, you know, experience bringing computer science to the medical space. And there are a lot of people that had already been working on this project. So, I had a lot of, you know, experienced people to learn from, which was really helpful. But then I also sort of started to become friends with people that were in sort of this larger space but working on different applications. So, people that were working with X-ray imaging, people that were working with robots, and I started as I learned more about this area, the more I found it interesting. And and then eventually, you know, understood sort of the broader world of it and wanted to stay in that space. And that's how I ended up with Phillips and.

Megan Benay

Did you make that the connection there while you were doing your doctorate?

Ayushi Sinha

While I was doing my postdoc actually, so I stayed in the same lab after my PhD for a postdoc. And actually, the event where I sort of met my future employers is actually coming up in March, I think. It’s called LCSR Robotics industry day. And I participated in that. My advisor encouraged me to make a, you know, 5-minute presentation. So, I I gave this, you know, short talk. And immediately after the talk, someone from Phillips approached me and said you should apply to Phillips. So, I did and they made an offer, so I decided to take it.

Megan Benay

OK, so moral of that story is don't be afraid to do presentations at conferences.

Ayushi Sinha

Exactly. The more visibility you can create for yourself and your work the better.

Megan Benay

I have nothing to contribute to a robotics conference, but it also sounds very cool. That sounds like a really cool conference.

Ayushi Sinha

Yeah, if you go on to the Homewood campus often, you should stop by it. I think it's open. And if you've been, if you walk by the, I think it's called the Robotorium. If you if you walk by Hackerman, the building, you know there's a sort of glass room with robots inside. It's usually there so you can walk through it and see people sort of talking about similar work, or maybe not similar, but sort of in the same broader scope of work that I did during my PhD. It's very interesting.

Megan Benay

That’s great. Well, thank you. We'll wrap with with our our one last question which is what inspires you right now?

Ayushi Sinha

I think a lot of things. Healthcare is such a, you know, sort of ever-present topic, right? It affects everybody. It affects you know people that are providing treatment and people that are receiving treatment. So, it's it's a really interesting field to be in just because there are so many different innovations happening both from sor of a automation perspective, but also just from techniques that are being sort of invented by physicians who, you know, decided that they don't want to open up people’s heads to provide treatment. They want to be able to provide this treatment with a small incision in the groin. I mean, there are people that we work with that invented certain guidewires and catheters that have helped bring, you know, their field forward. So, it's really interesting space to be in to be working with people that are innovating and to be innovating in that space. And obviously I think for almost anybody that is in the healthcare field to be able to make even the smallest difference in improving the care that patients are receiving and to potentially improve their long-term outcome, it's just a very, very fulfilling feeling and sometimes it's also a little bit frustrating just because you don't always see those impacts until sort of, you know, maybe years to come. So, it is. It is sometimes a little bit frustrating not to immediately see the effects, but then you sort of interact with the physician and you show them what you're working on and you see their reaction and you kind of hear comments from them on how that would help them provide better care. And it kind of brings back that, you know, confidence that eventually this will be in the hands of the people that will be helping patients feel better.

Megan Benay

It sounds certainly inspiring and like you're very fortunate because you're doing work that aligns with your why, your why for being in this field, your why for pursuing your doctorate in the 1st place and thank you for that reminder to keep being inspired by our whys.

Ayushi Sinha

Thank you so much. Yeah, thank you for giving me the opportunity to share.