Giant Robots Smashing Into Other Giant Robots

481: Oscar Health with Neetu Rajpal

June 29th, 2023

Neetu Rajpal is the CTO of Oscar Health. Oscar aims to make a healthier life accessible and affordable for all by refactoring healthcare to make great care cost less.

Chad talks to Neetu about working for a relatively new insurance company, reorganizing its structure by getting people into the right positions, and how incorporating large language models and generative AI is an inflection point that will help move things forward.

Become a Sponsor of Giant Robots!


CHAD: This is the Giant Robots Smashing Into Other Giant Robots Podcast, where we explore the design, development, and business of great products. I'm your host, Chad Pytel. And with me today is Neetu Rajpal, the CTO of Oscar Health. Neetu, thank you so much for joining me.

NEETU: It's great to be here. Thank you very much for having me.

CHAD: I want to talk about your role at Oscar Health, and your history, and everything that you've done. But everyone listening might not be familiar with Oscar Health. So let's start there; what is Oscar Health?

NEETU: Yes. Oscar Health is a health insurance company. We sell health insurance, primarily on the ACA marketplace across 20 different states in the U.S. We have just over a million members. So we basically sell health insurance to people.

One of the big, unique things about Oscar Health has been it's a very relatively new insurance company. So it's only been around for about ten years. And it was founded as a pretty standard tech startup. We've built all of the infrastructure for acquiring supporting members and providers, and brokers in-house. So we're fully cloud-native, distributed systems hosted on AWS and GCP with a giant data lake that supports all of our workflows. And this is a pretty unique integrated solution in terms of health insurance companies. So we're very much a tech-focused health insurance company. I've been at Oscar for about three and a half years.

I came to Oscar not from a healthcare background but just really mission-oriented and motivated to go help something in the healthcare space. I've spent most of my career building software, first at Microsoft and then at Conductor and WeWork. And I'm really excited to be here. It's really, really rewarding to be at a company that's serving primarily an underserved market in the ACA space.

CHAD: Well, I suppose full disclosure is in order. Oscar and thoughtbot have been working together for a long time now, actually, with Oscar as a client of ours. So I appreciate you joining us on the show, and I appreciate you working with us over the years. I think we worked...we started working with Oscar maybe when you were just in one or two states. And so, how have you handled that growth? And I think that's one of the complexities of the insurance space, right? Is every location is different in important ways.

NEETU: Yeah. Actually, it seems like Oscar and thoughtbot have worked longer than Oscar and myself.

CHAD: [laughs]

NEETU: So I think that's a pretty exciting, interesting statistic. And even during my time, it's been a great experience working with thoughtbot.

One of the big premise Oscar had was to build software that was segregated enough and isolated enough but composable enough that you could, in fact, bundle the full healthcare tech stack and then bundle it back together as you needed with configuration and scale it as you needed with, like, smart built-in scalability. And I think our ability to grow into multiple states and multiple counties has been, like, a good proof point of the fact that you can, in fact, do that.

In most cases, adding a new state is sometimes, at least from the tech perspective, from the software perspective is a combination of identifying the right configuration settings, mapping it to the features, and then just configuring the software to be able to support that new service area that we've just added. And that, I think, has been, like, a huge value add in terms of being able to add new locations to serve our members, add new providers, add new contracts. And our premise of building stuff or to unbundle and then bundle it back together as needed has really proven out a lot.

CHAD: I assume there have been some challenges along the way. What do you think have been among the biggest?

NEETU: I think the challenges have been an interesting combination of just learning the insurance business landscape then building software that aligns with it. So each county you might go into, each state you might go into, have its own set of regulatory requirements, have its own set of, like, reimbursement requirements, have its own set of, like, plan requirements, and these a new insurance company, I think along the way, we really did have to learn a lot of these things and sometimes by trial and error and that, you know, it's a pretty sometimes a long cycle with insurance. So that has been interesting and challenging.

CHAD: So not necessarily a technical problem but more, like, even just figuring out what it it should work.

NEETU: Yeah. I think definitely the business problem has been interesting on the technical side. So I started at Oscar about three and a half years ago. And my first questions as I was going through the interview process were, "Like, I don't understand why you built all of this stuff in-house, like, why aren't you just [inaudible 05:18]?" And those were rather naive questions to be asking a tech-focused insurance company.

Many years ago, when Oscar chose to use Elasticsearch, it wasn't a BAA-compliant solution. So you couldn't actually use a managed version of Elasticsearch. So we ended up hosting our own Elasticsearch cluster. We were one of the first or the first few people to actually sign a BAA with AWS a long time ago. So, technically speaking, I think the challenges have been around you don't always have the same set of tools available to you; at least, that used to be the case. This is rapidly declining rapidly. The availability of tools is almost the same now. And two, the ecosystem that you live in.

We still definitely have a service that has easy availability to a fax machine, and I think there are a few companies that are able to say that. But it's also the ecosystem that you live in, that you're trying to bring along with yourself, and also ecosystem you're using to support and build the tech stack. Both of those things have been rather interesting. In terms of actually building the software, that's, like, pretty standard regular set of software challenges.

CHAD: How big is the Oscar engineering team at this point?

NEETU: Oscar engineering team has fluctuated between 300 and 400 people over the past few years. I think we're about 350 or so right now.

CHAD: Generally speaking, how is the team organized to perform at that scale?

NEETU: I am actually a pretty big believer in first building a tech strategy that is tied to the business strategy before building the org structure. So, even at the Oscar engineering and tech team org structure, it has probably fluctuated quite a bit based on what the business needed. As of right now, we have dedicated folks that are aligned along individual set of audiences that we are supporting.

So we have a group of people who focuses primarily on the members and the member applications. We have a group of folks who focuses primarily on the provider and all the needs of the providers. And then we have a group of people maybe focused on just growth in general. And that could be growth through enrollments with brokers, enrollments without brokers, direct exchanges, or even some of our +Oscar business work.

So we've tried to, at this moment, align ourselves with the audience that we're serving. And, of course, like, no engineering team is purely vertical or purely horizontal. So we also have an infrastructure platform team that supports all of the areas.

CHAD: Do people sort of opt in, or are their roles advertised in each of those places? Or, as you reorganize into those structures, have you seen ways that work successfully in terms of getting people into the right position?

NEETU: Like, at the larger level in terms of the audience space, there tends to be pretty standard set of roles. As you might imagine, when the companies are smaller, there's a lot of investment or need for really, truly full-stack developers and full-stack engineers. As we've grown larger, there is a bit of specialization towards either the front-end portion of the house or the back-end portion of the house.

And we do have roles posted, and the roles some of them concentrate more towards the front-end tech. Some of them concentrate more towards distributed computing, and back end, and data engineering roles. And some concentrate more towards infrastructure engineering. And, as the teams always get bigger, there tend to also be, like, the supporting functions around technical product management and technical program management that are also part of the equation at the moment.

We do post all of these roles in those terms, with clarity around what the expectations are. As we re-org, we always prioritize internal candidates. We have a big enough team. People like to work on different set of problems and happy to align passions with our business needs because that tends to work out really, really well.

CHAD: Well, there's a reason why I'm asking about this particularly for you because one of the things that stood out to me about your background was the length of time that you were at Microsoft and the movement that you did between a variety of different roles across the organization. So that's an experience that you bring to the table. Was that something that you did intentionally?

NEETU: Yeah. I was at Microsoft for 18 years. And I think every couple of years, I did a new thing. And while I would like to very much say, yeah, I totally figured out that I wanted to be an engineer...

CHAD: [laughs]

NEETU: And then a product manager, and then, you know, an engineer again, or then a product manager again, or a leadership role again, or even the stack as in work on small APIs or work on full enterprise-grade products or cloud...I would like to very much say those were very thought through, you know, clearly pre-planned career moves; they were not.

And particularly, in my case, they were very much I chased problems. And I would get very obsessed or interested in a particular problem. And then, I would go dive deep into that problem and aim to go solve it. And by the time I had figured it out and solved it, there was always a new problem. And this was the wonderful promise of Microsoft where, you know, it is so large, and there are so many different ways to think about different problems, or different ways you can bring software to market, or problems in the world you can solve with it, that I particularly took full advantage of it.

And that hasn't really changed in terms of how the rest of my career post-Microsoft has gone, either. Every couple of years, I find myself in a space that is completely scary, completely new, completely different. For me, that tends to be my happy place, to be working on super difficult things that are very scary.

Mid-Roll Ad:

When starting a new project, we understand that you want to make the right choices in technology, features, and investment but that you don’t have all year to do extended research.

In just a few weeks, thoughtbot's Discovery Sprints deliver a user-centered product journey, a clickable prototype or Proof of Concept, and key market insights from focused user research. We’ll help you to identify the primary user flow, decide which framework should be used to bring it to life, and set a firm estimate on future development efforts.

Maximize impact and minimize risk with a validated roadmap for your new product. Get started at:

CHAD: It does sound like even though you might not have some grand plan, [chuckles] you are driven to seek out challenges.

NEETU: Yeah. I heard something really, really long time ago that kind of has stuck with me always, which is don't let your curiosity die. So, if you are curious, don't just let it shy away; just indulge in it. That's how you're always moving forward. For me, that's been a big theme as in, I really, really enjoy always learning. Always finding myself not to be the smartest person in the room is really, really great. And always moving forward and pushing hard, and solving bigger and bigger problems.

CHAD: So, when you were deciding to join Oscar Health, how concerned were you with the particulars of the individual position that you were applying for? Or was it more, like, I've identified this space and this problem that I want to be a part of?

NEETU: Yeah, that's a really, really great question. And I think it's one of the ones that when I mentor people, I tell them about as well, which is, I had a role with Conductor that was this head of R&D, CTO, and CPO, and a very traditional ladder style path to the next step or something along those lines or bigger. But I personally am not motivated by titles as much as I am about potential impact I can have.

My walk into healthcare out of MarTech or out of just, like, platform was very personal in terms of I have a healthcare story. Everybody really has a healthcare story. And I wanted to go do something in this space, utilize all of the skills that I already have, and then try to push the space forward.

And so, when I joined Oscar, I was not really that motivated by what the title was going to be. I was really looking for is, is there going to be an opportunity to have an impact? And is there going to be space for me to have impact? Am I going to be surrounded by a group of people who I can have impact with? That was the primary concern. And I did join Oscar as a VP of engineering. And that didn't last very long.

Within the next year and a half, I was already promoted to CTO in that role. And each one of those steps was very much motivated by will I actually be able to have impact? Will I be empowered to have impact? Was I equipped to have an impact? And was I going to learn a lot in that role? Those were my primary motivators. And that's always been the case.

It's also a hard-won lesson. Like, when I had my first kid, it was a really difficult time, personally. I had really bad postpartum depression. I had really hard time dealing with what is going to happen with my career, and I had to take a step back. And as somebody who was, like, super ambitious and wanted to go straight up, that was a really challenging thing to do. And that was a long time ago, and I've been able to rebuild, quote, unquote, "rebuild" my career back multiple times over.

So it's actually been, like, a hard-won lesson where it doesn't kind of matter what part of the rocket ship you have a seat on, as long as you have the seat on the rocket ship. And by rocket ship, I wish I was only talking about stock price; I'm not. I'm just talking about the ability to have impact. If I can be in this right group and I can have a voice, then it didn't matter what my title was.

CHAD: Yeah, thank you for sharing. And I think it feels like the pace of change accelerates. But in growing organizations, so much changes all of the time. So what you are doing...or the roles that exist will change as well. And the other thing...I often get asked by clients, "Should we be hiring a VP of engineering or a CTO?" I'm like, the definition of those roles varies so much between company, and team size, and everything. It's like; it really is difficult to answer that question because it is so different. It all depends on the needs of the organization.

NEETU: Yeah. I've done both of those roles, at least twice each time, twice now. And I've personally discovered that A, those roles don't really have a defined, like, there is no clear definition for either one of those roles.

The other thing that I think is more important is I don't think the organizations know what they want out of those roles most of the times, either. In my experience so far, VP of engineering is a little bit easier to define because it has a very heavy management component to it. And so I've ended up just defining those roles for myself every time, and so far, so good.

And the way I define those roles is very much, like, a CTO role is a half-business half-technical role. It's the role where you understand the business strategy and turn it into a tech strategy that supports the business strategy and pushes the business forward, and acts as the competitive advantage for the business. And your role is, in fact, to be the person who takes tech and applies it to pushing the business forward. That's one direction.

And the other direction, it is the role to take the business strategy and translate into technical terms so that you can bring and coalesce your whole team's motivation around it and bring the whole team around it. And I definitely find this very much at Oscar, where most people who join Oscar engineering and Oscar tech they all come motivated outside of Oscar already to help push the industry forward in one way or another.

So actually keeping that motivation in the right place all the time with authenticity and truth about what it is, how we're pushing the business forward. And helping the business move forward is an extremely valuable skill. So, from my perspective, the tech role is definitely half-business, half-tech with some variable sprinkling of management attached to it, and I think that's the CTO role. And the VP of engineering was very heavily bent towards management, I think, yeah.

CHAD: Well, so in that CTO role, we'll often be faced with understanding how industry changes, or new technology, or whatever can help the business. Obviously, we have a big one happening right now with artificial intelligence or large language models and machine learning. I have a couple of questions around this area. But, like, is there something in particular with that that you're either thinking about now in terms of Oscar or have already incorporated?

NEETU: Yeah. I think large language models and generative AI, in general, is definitely an inflection point. And I think it's an inflection point that is going to help push so many things forward. And I think healthcare is very interestingly poised towards pushing in that direction. There is everything from, like, shortage of providers or shortage of mental health providers to just a large availability of data in the healthcare space that can help discover things that may not be the easiest to discover through humans looking at it. So I think there's a huge opportunity.

I also think it needs, like, a bit of cautious evaluation. We are dealing with something that you can't break, so you can't move fast and break things when you're dealing with people. But there are interesting use cases where you could use generative AI and have it maybe not make a decision for a person but assist the provider to make a decision. Or you can use it in areas, at least in healthcare space, where there is a lot of inefficiency because of too much to do. And you can actually optimize a whole bunch of that.

So examples could be, you know, there's the standard examples of chatbots as in, is this doctor part of my network? Can I have an appointment with this doctor? Can I have an appointment and a prior auth with this doctor immediately? Like, those are, I think, a little bit more accessible use cases. And Oscar, also, we're exploring and figuring out which one of those makes most sense for us in our business. And there's tons of excitement and tons of, like, thought being put behind it. As of right now, I probably can't say all the things that we're working on at the moment. But yes, absolutely, this is a very big deal.

CHAD: So, speaking more generally around exploring new ideas, you mentioned at Conductor, you were in an R&D role, is there strategies that you've seen that work particularly well for doing research and development within an organization, exploring new ideas, figuring things out?

NEETU: Yeah. It definitely depends on which portion or which portion of the maturity stack you're on. If you're building software that is going to be sold to enterprises, you have to have some version of a promise of a date, and then you have to keep your promise. So, if you're selling enterprise software, you have to be predictable because your partners on the other side are relying on you.

And you can't actually get predictability without having enough stability in your team, and enough stability in your roadmaps, and enough stability in your tech environment [inaudible 22:30] So that you can figure out what you're going to build, you can figure out where you're going to build, and how you're going to build it. You can figure out what teams and resources you're going to allocate to it. And you can also have some confidence that when you're ready to send code out to production, there is some automated CI/CD system that is, like, going to help you make sure that the number of errors that you are sending out into the world is as few as possible.

And that superpower of being able to, like, churn out code and features and products like a machine on a pretty regular basis has the potential downside of not being able to disrupt yourself. And I think there are definitely a couple of strategies to do that. One of them is to allocate enough space by filling it with P2s as opposed to everything being a P0, that if you were to, like, displace the P2s with something that is new and available now, then you're going to be okay because you can live without the P2s. We had to have enough of those.

There is also the whole skunk works model, which I think works pretty well. I think it does need to be set up carefully so that you're not, like, destroying the motivation of everybody in the team. You do have the opportunity to do the skunkworks model. And, like, the generating ideas portion, we do this at Oscar, and I think this is also done elsewhere. We have pretty regular hackathons that are dedicated amounts of time, and we put them on the schedule. And, during that time, we just go explore and dream and go build other things. And we build it collaboratively with the rest of the business.

And there are definitely still, like, stories of things that were developed in a hackathon that served us really well for years on end. And I don't think that's ever going to change. But anything that is, like, a real solid enterprise production stuff probably needs a dedicated skunkworks something or the other so you can...I don't think it's a bad thing to have solid schedules. I think it's a really huge superpower to have solid schedules. I do think you have to have the discipline to be able to disrupt yourself; otherwise, somebody else will.

CHAD: Right. And so figuring out a structure, you can do that. And, like you said, you know, you don't want to ruin the morale or the excitement of the entire organization where people say, like, "Well, it's not my job," you know, or, "I wish I could work on this, but it's this team's." And you're not set up to capture that individual person's excitement over generative AI, for example, who is the one who's actually going to make it happen. And you're squashing that because they're not, quote, unquote, "supposed to work on it." I think that's a really difficult balance to strike.

NEETU: Yeah, I think it is. But I don't know; there are very few things in the healthcare space that anybody does that is easy.

CHAD: Yeah, that's a good point. Now that you've been in the role, well, at Oscar for a few years, what's next? What are the things that, like, you feel like are sort of motivating you each day you come to work?

NEETU: Oscar is in the ACA space, which serves primarily underserved communities. There have been, like, some recent examples of where people had to shutter doors in the ACA space. So the fact that we're still around and the fact that we're successful, like, we don't take that for granted. It's a really valuable thing to be able to survive in this market and to be able to grow.

We launched a mission-driven tech-focused company ten years ago based on the belief that you could build your own tech stack and be successful in a very tightly controlled market. And I think this is the year that we actually had to just, like, really prove out the use case. So, this year, Oscar is 100% focused on being insure co profitable. And once we finish proving that out, next year, we're going to prove out that it is... Holdco can be profitable, which is all of our tech stack, in addition to all of our insurance business.

And I'm very, very much looking forward to Oscar being successful this year in the mission, where I think it's just proving out that you can build a successful business and ACA-powered by your own tech stack. And not compromising on the outcomes for our members is really valuable and keeps me very motivated.

And then, looking at the future, looking at being able to be 100% full company profitable and potentially selling our software for others so they could also bring this efficiency to their businesses is extremely motivating for me. And, as of right now, this is what the whole company is focused on. And we're all super excited about it.

CHAD: Yeah, oftentimes, having a clear goal that everyone knows about and is working on can be really empowering for an organization, even if it's hard.

NEETU: Yeah. Like I said before, I think almost everybody I've interacted with at Oscar didn't come to Oscar, assuming it was going to be easy. They came to Oscar with full knowledge that it was going to be hard, but they were going to try anyway. So we are so close.

CHAD: Yeah. And, you know, that's one of the things that, at thoughtbot, we skew to working with clients that do positive things for the world, that deserve to exist for precisely this reason, is when you have a positive purpose in your organization, it makes working hard easier, [laughs] or solving big problems easier because you know that you're going to have a positive impact when you do.

NEETU: Yeah. The personal healthcare story was a motivating factor towards going to go do something that was, like, going to be, like, positive for the world. I wish I could say, yes, I'm going to solve all of healthcare's problems and [inaudible 29:02] not that way.

CHAD: Yeah. Well, that was actually what I was going to ask you about next is that's the thing about healthcare. And I've seen people almost get demoralized for it because you actually can't solve every problem. It's really difficult to solve all of healthcare. So, how do you sort of exist within that environment and not get demotivated?

NEETU: I think the more you dig into the problem, the more you realize how interconnected the way in, like, society, politics, money, power, and tech it really is. And it seems like if you try to solve the whole thing, it almost feels like you're going to boil the whole ocean. And it is definitely a blocker from even trying.

And, for me, it was just like, does that mean I stop? Do I apply the boy scout rule of, like, just trying to leave it better off, a whole lot better off, as much possible better off than I found it? I could do that. Or I could just go do something completely different. But there is no guarantee it's going to be as motivating or potentially as impacting. So, for me, it's very much about, like, if I could make even small changes, I know they'll be part of a greater whole. So I will just try to make those small changes because, alternatively, I may be neutral for the world at most.

CHAD: Well, Neetu, thank you so much for stopping by and sharing with us. I really appreciate it.

NEETU: Thank you very much for having me. It's been a great conversation.

CHAD: If folks want to get in touch with you or follow along, where are the best places for them to do that?

NEETU: I am easy to find on LinkedIn. I do have a Twitter account, but I'm not very active. It's @Nee2D2, and I look forward to hearing from people.

CHAD: Awesome.

You can subscribe to the show and find notes for this episode, along with a complete transcript, at If you have questions or comments, email us at And you could find me on Mastodon

This podcast is brought to you by thoughtbot and produced and edited by Mandy Moore. Thanks so much for listening, and see you next time.

ANNOUNCER: This podcast is brought to you by thoughtbot, your expert strategy, design, development, and product management partner. We bring digital products from idea to success and teach you how because we care. Learn more at

Support Giant Robots Smashing Into Other Giant Robots