Artwork for podcast Software Architecture Insights
From Finance to Healthcare: Navigating the Shift with Michi Kono
1st October 2025 • Software Architecture Insights • Lee Atchison
00:00:00 00:31:13

Share Episode

Shownotes

Michi Kono, the CTO of Garner Health, joins us to discuss the intersection of technology and healthcare. Garner Health leverages advanced data management techniques to help employers find the best healthcare providers for their employees. Michi shares insights from his extensive background in the financial sector, where he led significant tech initiatives, and explains how those experiences inform his current work in healthcare. We explore the complexities of the U.S. healthcare system, including data accessibility and the challenges posed by outdated practices. Michi emphasizes the importance of modernizing data systems and the role of startups in driving innovation in this vital industry.

The conversation with Michi Kono, CTO of Garner Health, delves into the intersection of technology and healthcare. Michi shares his transition from the financial sector, where he played significant roles at companies like Capital One and Stripe, to the healthcare field. At Garner Health, he leads initiatives that leverage data analysis to guide employees in choosing quality healthcare providers. The discussion highlights the importance of utilizing modern data management techniques to improve healthcare outcomes, underscoring how data-driven approaches can enhance the quality of life for millions. Michi explains how the complexities of the US healthcare system, including regulation and privacy concerns, present unique challenges compared to the financial industry. The episode also touches on the evolution of healthcare data management and the need for modernization within healthcare organizations to keep up with technological advancements.

Takeaways:

  • Michi Kono transitioned from financial technology to healthcare to make a meaningful impact on people's lives.
  • Garner Health utilizes data science to analyze medical outcomes and recommend quality healthcare providers.
  • The healthcare industry is complex and unique, reflecting challenges in data access and integration compared to finance.
  • Modern data management in healthcare is evolving, yet many systems still rely on outdated practices, slowing innovation.
  • Healthcare startups face unique challenges but have opportunities to leverage new technologies for improved patient outcomes.
  • Understanding data flow and machine learning basics is essential for future engineers in both healthcare and financial technology.

Links referenced in this episode:

Companies mentioned in this episode:

  • Gartner Health
  • Stripe
  • Meta
  • Capital One
  • Snowflake
  • Zocdoc
  • EPIC

Mentioned in this episode:

What do 160,000 of your peers have in common?

They've all boosted their skills and career prospects by taking one of my courses. Go to atchisonacademy.com.

Atchison Academy

How do you operate a modern organization at scale?

Read more in my O'Reilly Media book "Architecting for Scale", now in its second edition. http://architectingforscale.com

Architecting for Scale

Transcripts

Speaker A:

Hello and welcome to Software Architecture Insights, your go to resource for empowering software architects and aspiring professionals with the knowledge and tools they require to navigate the complex landscape of modern software design.

Speaker A:

My guest today is Michi Kono.

Speaker A:

Michi is the CTO of Gartner Health and a seasoned technology leader.

Speaker A:

At Stripe, Michi led teams overseeing core ML models for payment fraud prevention and identity verification.

Speaker A:

At Meta, he managed the global payment processing systems that powered all company wide commerce.

Speaker A:

At Capital One, he spearheaded the development of the company's inaugural machine learning engineering platform and registry.

Speaker A:

Known for scaling high performing distributed teams, Michie exceeded excels in data driven leadership and driving transformative engineering initiatives.

Speaker A:

Michi, it's truly a pleasure to have you here and welcome to Software Architecture Insights.

Speaker B:

Thank you for having me.

Speaker A:

Well, you've got quite a resume and you've been involved in the financial aspects of technology, the financial space for quite some time.

Speaker A:

But now you've moved, you've moved into the healthcare industry.

Speaker A:

What led to that change for you?

Speaker B:

You know, I thought that first of all, from a skill set perspective, you know, both of the industries are regulated.

Speaker B:

So I thought that my background, being familiar with that type of problem would be helpful here.

Speaker B:

But I think primarily for healthcare, you know, and especially at this startup, it was nice to work on a problem that I felt ever really truly matters to people.

Speaker B:

It's if we're successful in our mission, then I think we'll make a real dent in the quality of lives for millions of Americans.

Speaker B:

And so that was the main driver for joining this company, Garner Health.

Speaker A:

We're not going to go into a lot of detail about Garner Health specifically, but just to kind of set the groundwork for the listener, can you tell me a little bit about what Gartner Health does?

Speaker A:

Just your nutshell, elevator pitch, if you will.

Speaker B:

Yeah, we use big data analysis, data science to analyze the medical outcomes, the claims data of patients across the country to determine which doctors, which providers are effectively good or bad.

Speaker B:

And we take that data and give that list of the best doctors in the country to employees as an employee benefit, we sell it to employers as an employee benefit and so they can then use that to find the best care.

Speaker A:

So your specific focus on essentially finding quality doctors for patients, that's your specific.

Speaker B:

That's right.

Speaker B:

Yeah, that's right.

Speaker A:

Now, healthcare is one of the most complex industries around.

Speaker A:

There's regulation, there's privacy concerns, some complex interrelationships between insurance companies, hospitals, doctors, patients, the government, and obviously there's lots of money involved.

Speaker A:

As you Mentioned earlier, a lot of the same things that happen in the finance industry.

Speaker A:

There's also one other thing that makes it different, I think from the finance industry and that's there's massive differences from country to country.

Speaker A:

You know, US Healthcare is different than every other industry.

Speaker A:

I'm assuming you focus primarily on the US healthcare industry.

Speaker B:

Yes, I think that the US healthcare industry is actually quite unique in the world in some ways because of its complexity that we built on top of it.

Speaker B:

And so our startup really works specifically in the US because of that dynamic.

Speaker B:

But the problem of finding great doctors is probably universal.

Speaker B:

And funny enough, because of the insurance system that we have here, there's actual an ability to actually analyze that data and figure that out in a way that's unique to the market.

Speaker A:

Technology for data management in healthcare industry is relatively new for healthcare.

Speaker A:

I mean, it's been within the last decade or so that we've seen a major increase in the use of data and data processing for the healthcare industry.

Speaker A:

How has healthcare data management changed over the last decade or two?

Speaker B:

Well, I think that depends, like with any company, you know, depends on the company itself.

Speaker B:

But I think the, I think there's more tools available, more solutions available.

Speaker B:

Whether, you know, for example, many companies are starting to go into the cloud.

Speaker B:

The healthcare industry is probably a little bit of a late mover in that as a whole.

Speaker B:

And with that there are tools that weren't, that didn't exist.

Speaker B:

You know, I think a really sort of prominent example is like Snowflake didn't exist really as a company.

Speaker B:

They IPO, you know, less than 10 years ago.

Speaker B:

And so those types of things make data, data management a little bit easier to scale.

Speaker B:

And you know, the idea, the term data management is actually a loaded term.

Speaker B:

So I didn't know if I meant to use it the way I just did.

Speaker B:

But that concept of data management where you're trying to govern the data, make sure it's being used correctly, you're violating any laws that it's being privacy first.

Speaker B:

Like these types of concepts, doing that at scale is really hard.

Speaker B:

And so in the old days you just had a big database and lock it all down.

Speaker B:

Now you need to do all this distributed processing and put it all over the place and use it for all these different use cases.

Speaker B:

And so the technology around that's definitely evolved.

Speaker B:

And I think the industry is still early in adopting that end to end.

Speaker A:

Yeah, so scaling is actually a huge issue for you.

Speaker A:

I mean there's 300 million people in the US and every single one of them has some sort of health care concern at some level.

Speaker A:

And so you're talking about billions of data transactions a year that you have to, that technology companies have to deal with.

Speaker A:

How closely tied to that is your company?

Speaker A:

Or are you more peripheral to that main data processing action that goes on?

Speaker B:

I mean, we do ingest data without getting into specifics of how much.

Speaker B:

But I will say that the healthcare space in some ways is maybe there's some aspects that's easier and some aspects that's harder.

Speaker B:

So it's easier in that the volume of data that you're involved with is actually less.

Speaker B:

Like there's way more financial transactions per person, if you think about it.

Speaker B:

And they're not as real time.

Speaker B:

If you think of a stock trade or something, or a payment, that has to happen 100 milliseconds, decision fraud, whatever has to be accounted for.

Speaker B:

Whereas with an insurance claim or something, you've got some time.

Speaker B:

And maybe there's some parts of that that are real time, like whether or not to approve a claim, doctor's visit or something.

Speaker B:

But that stuff is easier.

Speaker B:

I think there's other parts that are harder in that not all the data that you need is necessarily actually available, connected to the Internet and ready to go.

Speaker B:

You know, medical doctor notes are on paper or whatever it might be.

Speaker B:

So there's all these other factors that I think make it hard.

Speaker B:

Whereas in the financial world, like almost everything is at this point is digitized.

Speaker B:

So right.

Speaker B:

Like at the beginning of the credit card era, they would, you know, use that little thing to like press the credit card onto the ink paper.

Speaker B:

So they're past that.

Speaker B:

But if you think about it, the medical industry is still in that world today.

Speaker A:

Right.

Speaker B:

So that's the trade offs.

Speaker B:

And I think that can also basically slow down the adoption of fully end to end modern tools.

Speaker A:

So is the problem with that data access the fact that doctors are still using paper?

Speaker A:

Or is it that, for example, or is it that doctors are using systems that aren't connected to each other?

Speaker A:

Or is it both?

Speaker B:

It's definitely both.

Speaker B:

There's a question mark about how much doctor's notes, for example, are necessary for all the transactions that have to happen around health insurance things.

Speaker B:

But I think it's a treasure trove of data to use to understand what the doctor was thinking and whether it was correct or not.

Speaker B:

So part of it is that yes, the doctors are using sort of offline systems.

Speaker B:

And that's been okay because that hasn't blocked the right outcomes from occurring up until now.

Speaker B:

And still now, but obviously there's a treasure trove of data around that stuff, around what they're writing down and you know, and transcribing that into digital systems.

Speaker B:

But there's another aspect which is that you can operate a doctor's office and be pretty much almost all in offline.

Speaker B:

Right.

Speaker B:

So in fact that's a big problem with things like doctors availability and booking where, you know, there, there are doctors, we have to call the office to even know if they're available.

Speaker B:

So there's definitely a heterogeneity at the edges.

Speaker B:

This is kind of, I guess equivalent to like mom and pop restaurants.

Speaker B:

They can just be cash based and you would never.

Speaker B:

They can have paper menus or they can use square and you can order everything digitally and you can get McDonald's and do the kiosk.

Speaker B:

And so if you're trying to do like, if you're trying to design a great menu or something, McDonald's has all the data about what's going across, going on across all their franchisees.

Speaker B:

But then if you have like a local mom and pop, that data is just sitting in a, I mean if you're lucky, there's a pile of receipts.

Speaker B:

Right.

Speaker B:

So I think that dynamic is definitely a part of what makes it hard in healthcare and that there isn't a central common way that everyone does it.

Speaker B:

Unlike in financial money movement.

Speaker A:

Do you see that as improving or is this a problem that's, you know, nothing's going to change.

Speaker A:

A Ma and PA restaurant that hasn't changed so far to, you know, have a kiosk based system that's always going to be there.

Speaker A:

But do you see the healthcare industry is actually moving towards automation.

Speaker A:

Are they incentivized to do so or are you always going to have the doctor that's doing something on paper because they've done it that way for 100 years?

Speaker B:

I think it depends on the activity.

Speaker B:

So for example, accepting insurance is implicitly a centralized system.

Speaker B:

Payments is not getting reimbursed by healthcare insurance provider is.

Speaker B:

And so there are some activities that are probably more centralized and ripe for innovation and modernization.

Speaker B:

And then what's the problem on the other side of that is like the, the way that you integrate, let's say you're an insurance carrier, like the process in which the rails in which you get that claim and reimburse the patient or however it's done, or pay the doctor and bill them, like all that stuff is, you're still dealing with this mom and pop at the end of the day.

Speaker B:

So we have more and more centralization of providers and hospitals and that that actually kind of helps move things forward.

Speaker B:

But you'll always have this edge case of long tail doctors who aren't.

Speaker B:

So there's two parts that are maybe ripe for innovation in the space.

Speaker B:

One is obviously modernizing these edges, but I think also the big players in this space, of course they're working hard to adopt the cloud and move to more modern systems and enable ML and do all this stuff.

Speaker B:

And they're probably discovering as they're thinking about doing ML or even data science that hey, we have all this data and we didn't realize we needed it for that use case.

Speaker B:

And now the systems that we want to, that the data's sitting in is 25 databases across the company and all these different formats.

Speaker B:

So we can't actually run even a simple query across those databases.

Speaker B:

Like those are the kinds of problems we're having, let alone access rights and can I even use it for that?

Speaker B:

And like the person that collected this data, that person left the company, we have no idea what was agreed to.

Speaker B:

There's all these problems around governance as well.

Speaker B:

And I think that's probably, that's the other end of the spectrum that's probably happening.

Speaker A:

Imagine the MANPA doctor's office is probably being forced into modernization, probably much more so than the Ma and PA restaurant is because of these issues.

Speaker A:

Almost all medical transactions deal with some form of insurance or Medicare or something like that.

Speaker A:

So there's regulation involved at all levels and that's forcing a level of modernization that they may not normally be comfortable with.

Speaker A:

Is that a fair statement?

Speaker B:

I mean, I think that all of, I mean the mom and pop diner is too, when they start to accept a credit card, I mean, so I do think that that is happening.

Speaker B:

I think the really sort of modern industry, big data data processing, like how do you secure data?

Speaker B:

How do you share data?

Speaker B:

Those questions are not being managed by those like small physicians offices.

Speaker B:

Right?

Speaker B:

They're at best they're a stakeholder and maybe not even aware that that's happening.

Speaker B:

And I think that's where the transformation's happening.

Speaker B:

And then meanwhile there's smaller startups that are trying stuff out and there's startups like zocdoc and there's EPIC and stuff trying to bring tools to those providers.

Speaker A:

And is that fundamentally different than the financial industry?

Speaker A:

Is it just that healthcare is a decade behind but catching up or is it that there's something fundamental that's different?

Speaker B:

I think that the fundamental difference might be that for the most part, financial services in particular is in many ways a game about data movement.

Speaker B:

Like in its core, yes, you can set handsome on a dollar, but most of the time, like, you know, it's, it's more about like I owe you money or you owe me money and then we make that hole by moving money later.

Speaker B:

And so if you think about it like a bank, yes, back in the day they had a big vault and they went back there and picked up the money and gave it to you.

Speaker B:

But in practice today it's just zeros and ones in the database and then they reconcile that once a day or once a month or whatever it might be.

Speaker B:

And so in that way, the financial industry is much more motivated to modernize and be real time and be really good at their data systems in a way that maybe drove them to adopt these technologies like cloud and stuff much sooner than other regulated spaces because they kind of have to like it is what they actually do.

Speaker A:

So the financial industry is required to because it's the way they move money, it's the way they do their business is that way.

Speaker A:

But that's not true in the healthcare space.

Speaker A:

And so in healthcare space it's a value much more of a value add versus a core part of the business.

Speaker B:

Like the data.

Speaker B:

In financial services, a lot of times the data is the value being moved around.

Speaker B:

Right.

Speaker B:

when you get if I venmo you $:

Speaker B:

And in healthcare, the evaluation of what happened and the record keeping in or determining the best doctor is like these are, these are layered on top of the industry.

Speaker B:

And so yes, you have to protect that data.

Speaker B:

And so there is an aspect to it, but because it's sort of a second order, like the actual service is the medical care, it's second order to support the medical care.

Speaker B:

There isn't as strong an incentive to move quickly.

Speaker B:

So they're going to move more carefully once it's proven out.

Speaker B:

And so I'm sure it's good to see these banks of stuff moving all in on the cloud.

Speaker B:

It's good to see, I mean hedge funds have been on doing that stuff for a while, but like there's these other industries that are doing that well.

Speaker B:

And so that can make it obvious that okay, healthcare could probably get there too.

Speaker B:

And so that's why I think it's slightly delayed and there's some lag between where all the startups and all the tech companies went and then where the healthcare Companies are following now.

Speaker A:

So the issue isn't that healthcare is behind.

Speaker A:

The issue is that healthcare has been much more deliberate because they've been able to be more deliberate.

Speaker B:

I mean, that's probably, yeah, it could be phrased that way.

Speaker B:

I'm sure there's a little bit of both of that you don't want to.

Speaker B:

I'll say this as an engineering leader, like, I don't want to be in a world where I'm saying, I've got this cool solution, how do I use it?

Speaker B:

Ideally, you've got a problem you want to solve and so there's some problem that needs to be solved and technology is the best answer, then you would mitigate that by doing it.

Speaker B:

And so there's a world where you can get by with what was in place and at some point it breaks down.

Speaker B:

And so once we start doing like, okay, we want to, you know, we're a pharmaceutical company, we want to do agentic research on our data sets.

Speaker B:

Wait, hold on.

Speaker B:

Our database is not ready for that.

Speaker B:

It's not built for that or not.

Speaker B:

We don't have the lineage in the catalog to understand our data.

Speaker B:

So we can't do it.

Speaker B:

Right.

Speaker B:

And so that would then force the conversation.

Speaker B:

Oh, maybe we need to think about how we're actually storing our data, how do we process it?

Speaker B:

And that's going to then trigger the upgrades and stuff you might need.

Speaker A:

So that's how that works with the technology.

Speaker A:

Let's talk about the people aspects.

Speaker A:

And by the people aspects, I'm not talking about the doctors, I'm talking about the engineering teams.

Speaker A:

Now when the financial sector went through the modernization process, the engineering teams went through a modernization process as well.

Speaker A:

Right.

Speaker A:

You're moving from older big mainframes to cloud based computing systems and processes that changed to go along with that.

Speaker A:

And the engineering teams had to adjust and grow and modernize themselves, both their skill sets, but also their culture in order to make that happen, as you say.

Speaker A:

I think a lot of that was inevitable because of the industry itself.

Speaker A:

How is that working within the healthcare industry?

Speaker A:

Do you have the same sort of of team modernization process that has to go on and is going on?

Speaker B:

So yeah.

Speaker B:

So first of all, I observed some of that firsthand when I spent time at Capital One.

Speaker B:

They had just declared that we're going to go into the cloud.

Speaker B:

And there was a rush sort of inside the company to adopt those skill sets.

Speaker B:

And the company provided support and training.

Speaker B:

So not everyone made that jump.

Speaker B:

And some people left, but the majority, I'd say, adopted That I can't speak for this industry as a whole, healthcare industry as a whole.

Speaker B:

I can only imagine there's a similar thing going on, company by company where someone says, hey, there's this AI thing, it's really cool.

Speaker B:

I heard we can use it for research.

Speaker B:

How do we do that?

Speaker B:

And then no one knows how.

Speaker B:

And so there's this, okay, let's make a lab and let's get a lab team and have them go research this.

Speaker B:

So I'm sure that's happening and upskilling is happening.

Speaker B:

I think it's an interesting time for startups obviously because these types of like the industry run data tooling has gotten so good in the last like less than 10 years that a company like Gartner doing all this big data analysis is much more accessible than it ever was.

Speaker B:

And, and you can build things the right way as you go.

Speaker B:

So as a startup we're hiring people from day one that use all the modern technologies and tools that are in the industry and so we're not impeded by that.

Speaker B:

But I do imagine that the bigger companies with thousands of employees, like, yeah, there's probably a multi year journey, it's probably going to take them half a decade to move most of the culture, engineering culture to the next thing.

Speaker B:

Because that's about how long it took for the financial industry as well.

Speaker A:

That's an interesting take.

Speaker A:

I hadn't thought of it that way before, but I suppose if you think about it, the financial industry is kind of bootstrapping itself, it's growing itself from within.

Speaker A:

Capital One is changing its culture or has changed its culture.

Speaker A:

And same thing is true of course, with all of the major chase, all the major players, I think we all in the technology industry, we think about Capital One because we're kind of a major player, especially in the cloud and growth of the cloud, but everyone's doing that.

Speaker A:

But it's the industry itself that's self growing.

Speaker A:

But in healthcare, what it sounds like, and please correct me if this is wrong, I'm going to take a step back for a second.

Speaker A:

When I talk about healthcare technology, I'm referring to the data back end systems that we're talking about.

Speaker A:

I'm not talking about things like patient care, AI monitoring systems and the patient side of it.

Speaker A:

I'm talking about the data management side that we're talking about.

Speaker A:

But I imagine on a healthcare side it's much less the industry modernizing itself as much as it is startups like Gartner coming in and building modern systems that solve problems for the industry that they start adopting, was that an incorrect assessment or do you see, is that something else that's different than from the financial industry?

Speaker B:

I would give more credit to the big players that are there.

Speaker B:

I'm sure, I'm positive that there is strong work inside those companies trying to, to modernize and replace their old data centers with databricks or snowflake and bring on modern data governance practices and whatnot.

Speaker B:

I just know from watching it happen at Capital One, it's going to take years.

Speaker B:

You can put the tool in place and then go, okay, cool, we have 100 databases that have to be migrated over to this whatever lake we've chosen, let's say use databricks.

Speaker B:

We have, you know, 1,000 data streams coming in that no one's touched in 10 years that we have to go now.

Speaker B:

You know, we have to go wake up somebody and figure out how that was written.

Speaker B:

Right.

Speaker B:

And so there's all this data that's there.

Speaker B:

We've never actually sat down and categorized what exactly it's for, or what the valid values are, or whether the null value distribution is correct or not, or what the sanity check should be like.

Speaker B:

None of that's there.

Speaker B:

And so it's going to take them concerted effort, years to migrate all that stuff.

Speaker B:

And then once you have it in one place, you can start to do the really interesting work.

Speaker B:

But you can imagine a world where you're trying to figure out the quality of care in the country and you are an insurance carrier that's purchased other insurance carriers over the years.

Speaker B:

And so you've got historical patient data in 10, 20 different systems that you've never really had to migrate before.

Speaker B:

And you kind of did this analysis as you needed it.

Speaker B:

But now you want to do lots and lots of machine learning experimentation on that data as you're trying to get it in one place.

Speaker B:

And then you realize, wait a second, we are storing this data in all kinds of crazy ways that we hadn't thought about before.

Speaker B:

And now that we're trying to put it together in one place, it's falling apart.

Speaker B:

And so, for example, this data set over here didn't collect gender.

Speaker B:

And so is it okay for us to do a model on this data and train on it when maybe there's a bias in this data that isn't reflected?

Speaker B:

Maybe it's actually mostly males?

Speaker B:

We have no idea.

Speaker B:

And so because you can't get the providence of that data, it can harm your ability to do big data scale data research on this stuff.

Speaker B:

So those are the Kinds of problems that you encounter.

Speaker B:

And so then you have to go all the way back to the source and be like, what was this source?

Speaker B:

How do we get it?

Speaker B:

Where did it come from?

Speaker B:

Has it been changed?

Speaker B:

Maybe someone at some point assumes stuff and changed stuff in there and you'll never know.

Speaker B:

So those are the kinds of problems you deal with.

Speaker B:

And it just takes a long time to go through that.

Speaker A:

The tech industry as a whole right now is not growing like it used to.

Speaker A:

Right.

Speaker A:

You know, many people blame the growth of AI for the slowdown in the tech industry.

Speaker A:

Many people blame financial, political, economic reasons.

Speaker A:

Whatever the reasons are, tech industry really isn't growing like it used to.

Speaker A:

Is the same thing happening in healthcare technology or is healthcare technology, because of this modernization that's going on still a growing industry?

Speaker A:

Growing fat?

Speaker A:

It's growing, but growing faster than technology itself.

Speaker B:

I'll be honest to say.

Speaker B:

I don't know if there's statistics on that specific sector or not.

Speaker B:

I do think it's hard to.

Speaker B:

It's hard to be a startup that's not doing AI right now.

Speaker B:

I think you're probably.

Speaker B:

But if you're doing AI, perhaps it could be easier.

Speaker B:

But I think my sense is that the white collar career, sort of engineering career in general is just harder.

Speaker B:

And there is a lot of, there's an interest in the industry to solve that problem, which is why our company exists.

Speaker B:

But I don't know if it's, you know, if there's a specific healthcare engineering sector decline or not.

Speaker B:

I'm not 100% sure that's fair enough.

Speaker A:

Yeah.

Speaker A:

I can tell you by the way too that even in AI space it's hard to grow right now because there's so many people in the AI space.

Speaker A:

That's true.

Speaker A:

Yeah.

Speaker A:

I have an AI startup that I co founded with a friend of mine and just getting people interested is hard.

Speaker A:

Not because people aren't interested in AI, it's because which of these other hundred companies are doing the exact same thing and how do you fit relative to that?

Speaker A:

So it is an industry wide thing from that standpoint.

Speaker A:

So what about non US healthcare?

Speaker A:

Is there opportunities there like there is in the US healthcare or is this truly just a US healthcare thing?

Speaker B:

I think it's our company hasn't looked into going beyond the us.

Speaker B:

I think there's a plenty large market here.

Speaker B:

But the US market is uniquely inefficient because we as patients have no idea what the cost of healthcare is.

Speaker B:

Your employers have a sense of health, insurance carriers do.

Speaker B:

But you have no idea.

Speaker B:

And oftentimes even if you did know, you don't care.

Speaker B:

And so that dynamic is unique.

Speaker B:

Also we're not single payer.

Speaker B:

And so there's some interesting things about our system that's different.

Speaker B:

And that unique inefficiency creates then suboptimal outcomes for patients.

Speaker B:

Just to give you an example of what I mean by suboptimal.

Speaker B:

This is not malicious intent, but there's a world where the doctor who gives you a, a prescription for something, they're incentivized to have you come back ten more times, right.

Speaker B:

Because they get paid for each visit.

Speaker B:

So even if nothing is going on, if they can just get you to come back, they make money and insurance carriers and everyone else makes money.

Speaker B:

And so as long as you're willing to do the time, you'll come back.

Speaker B:

And that means you can say the same thing for unnecessary treatments, unnecessary surgeries, whatever it might be.

Speaker B:

And so that's the inefficiency that's unique to our market.

Speaker B:

And it can create situations where inadvertently, hopefully you can end up in a path where you're seeing a whole bunch of doctors and bouncing around between referrals and having all these follow up treatments even when you don't need it.

Speaker A:

Having been in the financial industry and now in the healthcare industry, what advice can you give to someone either new in the technology field or maybe looking for career growth within the technology field?

Speaker A:

What advice can you give them as far as the value of a career in healthcare technology versus financial technology versus any other technology?

Speaker B:

I think the, you know, I think back to like engineering is a constantly evolving field.

Speaker B:

ooks very different than like:

Speaker B:

Right.

Speaker B:

And I think about the trends, the decade to decade trends.

Speaker B:

So for example in:

Speaker B:

Like that's when the iPhones and the Android phones got really big.

Speaker B:

And so all that stuff where they did product engineering, I think that's where the really hot, really fast movement of technologies occurred.

Speaker B:

And at the beginning of:

Speaker B:

so now we're Fast forward to:

Speaker B:

But I think generalizing it a little bit further, you could say that data related engineering is I think the hot topic of the decade, we're seeing tons of new database technologies coming out.

Speaker B:

There's the whole thing with vector stores and Pinecone.

Speaker B:

That company went from nothing to a huge company overnight.

Speaker B:

And hype aside, it's here to stay as a category of technologies.

Speaker B:

And so the underpinning systems behind ML, in particular AI and ML, and the systems that then manage that data well will, I think, be the core things.

Speaker B:

And so if I'm talking to giving advice to people, I would say that as much as possible, try to understand the basics of what's involved in a really simple ML algorithm and how you would implement that and how you would train a model and how inference works.

Speaker B:

I think just having that basic understanding, understanding of data and how it flows through a system is probably the most important skill set for the next decade.

Speaker A:

Great.

Speaker A:

Thank you.

Speaker A:

So my guest today has been Michi Kono.

Speaker A:

Michi is the CTO of Garner Health and a major player in the data architecture technology involved in managing the US Healthcare industry.

Speaker A:

Michi, thank you so much for joining me on Software Architecture Insights.

Speaker B:

Pleasure to be here.

Speaker B:

Thank you.

Speaker A:

Thank you for joining us on Software Architecture Insights.

Speaker A:

If you found this episode interesting, please tell your friends and colleagues you can listen to Software Architecture Insights on all of the major podcast platforms.

Speaker A:

And if you want more from me, take a look at some of my many articles@softwarearchitectureinsights.com and while you're there joining the 2,000 people who have subscribed to my newsletter, so you always get my latest content as soon as it is available.

Speaker A:

Thank you for listening to Software Architecture Insights.

Chapters

Video

More from YouTube