Back to Podcast
A/B Testing
Feature Flags
ROI

How FLYR helped an airline grow revenue by double digits

S2 | E13
Nov 20, 2026

Summary

Airline seats are perishable inventory, and small decisions about pricing and product offers can have enormous consequences. Vamsi Kodam, Senior Director of Product Management at FLYR, joins host Ashley Stirrup on The Experimentation Edge to explain how FLYR is helping airlines modernize commerce, personalization, pricing, and offer orchestration.

Vamsi revisits a formative "human versus machine" experiment in airline revenue management: total revenue barely changed, but looking across the network revealed where human overrides helped, did nothing, or actively hurt performance. He also shares how FLYR applied multi-armed bandit experimentation to an airline pricing problem and saw double-digit revenue improvement, why he studies other practitioners' research before designing tests, and how airline experimentation can move beyond price to personalized product bundles.

📝 Read the full blog post →

Chapters

00:00 Meet Vamsi Kodam of FLYR
00:17 Building a modern airline commerce platform
01:37 Riyadh Air and a cart-style booking experience
02:28 Product catalog, pricing, and offer orchestration
03:37 Vamsi's first A/B tests in airline pricing
04:23 Experimentation beyond e-commerce
05:28 When human pricing overrides help or hurt
07:11 Why every experiment teaches you something
08:23 The lingering human price override
10:22 Multi-armed bandits and double-digit revenue improvement
12:01 Building experimentation into airline products
13:16 Learning from research and other industries
15:16 AI-assisted research monitoring
15:39 Experimenting beyond price and into product personalization
17:27 AI and faster product development
18:18 Feature flags as a safety net

Notable Quotes

“There's nothing like a failed experiment, I would say. There's always a learning.”

“Connecting that intuition to the data can help.”

“With a multi-arm bandit, the advantage is that you don't have to wait for the experiment to conclude.”

“The airline has seen double-digit revenue improvement.”

“There's a lot to pick from what others are doing, and you don't have to do everything on your own.”

Transcript

The Experimentation Edge - Vamsi Kodam

Vamsi Kodam (00:00) in the last six to twelve months, the the biggest benefit as a product person like from LLM or the AI technology is around the speed of development. So some of the things that used to take, let's say, a year or two and are done in like a quarter or in the past, like what I used to think about a year is l is about a quarter now. And

Ashley Stirrup (00:18) Yeah.

Vamsi Kodam (00:18) Things are moving fast and so that gave us like opportunities to consolidate, like a lot address a lot of tech debt more proactively and also pick on a lot of features that could have taken much longer. So we were able to move faster

Ashley Stirrup (00:31) Yeah.

Vamsi Kodam (00:31) on those,

Ashley Stirrup (01:04) Hello and welcome to today's episode. I'm excited to have Vamsi Kotem, Senior

Ashley Stirrup (01:09) Director of Product Management at FLYR. Welcome to the show, Vamsi.

Vamsi Kodam (01:13) Hey, thank you, Ashley. great to be on the show.

Ashley Stirrup (01:16) Yeah, I'm excited to have this conversation. why don't we start things off by you telling us a little bit about FLYR?

Vamsi Kodam (01:22) FLYR is a Series D travel technology company building a modern commerce platform for airlines. And if you look at the the technology that airlines use, a lot of it is built many decades ago I would even say a good part of it is pre internet age and FLYR is completely building it for like the modern era to create

Vamsi Kodam (01:42) a consumer centric like personalized experience for airlines so that they can cater better to their customers.

Ashley Stirrup (01:50) Got it. And are your primary users the airlines themselves or their customers?

Vamsi Kodam (01:56) It's a airline user. So this is a enterprise product, so I would say this is a B to B to C. So we provide this product

Ashley Stirrup (02:02) Got it.

Vamsi Kodam (02:02) and airlines use this for responding to their customers.

Ashley Stirrup (02:07) Got it, got it. it's funny as somebody's been in the software industry for a long time, I'm always sitting there at the airline counter watching the number of clicks that somebody has to do to move me from seat A to seat C. And I'm like, dang, that's a lot of clicks, So I'm sure there's a

Vamsi Kodam (02:24) Mm-hmm.

Ashley Stirrup (02:24) a lot of opportunity to make airline employees a lot happier with with FLYR. So yeah, yeah.

Vamsi Kodam (02:29) absolutely. Yeah. And some of the

Vamsi Kodam (02:32) technology that we are building we have had some very good feedback from the early adopter of the technology.

Ashley Stirrup (02:38) Got it. And can you tell us a little bit about some of the early adopters?

Vamsi Kodam (02:42) so we have we are working with Riyadh Air the new flagship airline that the Saudi government has launched and instead of using the decades old we call it like PSS, like the passenger service system, so that technology they adopted the FLYR's off-road platform for catering to their customers.

Vamsi Kodam (03:01) And for example, one of the differences on their platform, if you look at it, is they have a a cart experience. For example, when you're on a airline website, you go and shop and then you come back and it doesn't have the context of what you looking for. Whereas here you go and you create it and you put it in the cart, like what you do with Amazon or any of the e-commerce companies that you work with. And

Ashley Stirrup (03:22) Yeah.

Vamsi Kodam (03:23) So it's yeah, you can c put them in the cart and come back to it and it understands your preferences over the time and so on.

Ashley Stirrup (03:29) interesting.

Ashley Stirrup (03:30) And can you tell us a little bit about your role there?

Vamsi Kodam (03:33) Yeah, my role is I'm a senior director of product like for our offer and optimization suite of products and this is everything around the product catalog where we define what are all the products that the airline is able to sell. Obviously the airline is selling the flight to go from A to B, but beyond that there are like a lot of other products that the airlines has and the so that the product catalog and then how you price them. There's a lot of

Vamsi Kodam (03:59) nuance and complexity around pricing. So the the pricing engine and then there's a offer orchestration that happens so which takes into account like every request and looks at the product catalog and pricing engine to create an a personalized response and does it in milliseconds

Ashley Stirrup (04:16) Yeah, yeah, I can't help but as I'm listening to you think boy there's a lot of complicated stuff going on in the background for something like the airline industry with such a

Vamsi Kodam (04:23) Mm-hmm. Mm-hmm.

Ashley Stirrup (04:25) perishable asset, right, as a as a seat on a flight and everything, right?

Vamsi Kodam (04:27) Mm-hmm. Mm-hmm. Mm-hmm.

Vamsi Kodam (04:29) Yes, yes. So there's a yeah, it's perishable inventory. You got it right. And

Ashley Stirrup (04:33) Yeah.

Vamsi Kodam (04:33) yeah, so once the flight takes off, then there's no you can't monetize that. Yeah. Yeah, yeah. You can't sell the same to price, yes.

Ashley Stirrup (04:36) That's right. That seat is gone. And you can't sell the same seat twice. So yeah,

Ashley Stirrup (04:41) when did you first start doing A/B testing?

Vamsi Kodam (04:44) would say I started as a data scientist and I really loved l doing the math problems and like kind of apply it to a business context and see the outcomes come out of it. when I first started

Vamsi Kodam (04:56) I was working with an airline and they use these decision making platforms to primarily for determining like at what price levels they should sell. And this is something that they were doing like they do automated systems at scale like for take millions of decisions every day. And I

Ashley Stirrup (05:13) Right.

Vamsi Kodam (05:14) I was working with an airline and that was

Vamsi Kodam (05:16) looking for how to improve that automation. And I worked on A/B tests for them in my first year like this is like almost like fifteen, sixteen years ago, I was I did my first A/B test

Ashley Stirrup (05:27) How do you think about A/B testing in a world like yours? Because obviously you don't have, you know, the volume of an e-commerce website. You're still relatively new, you're probably introducing a lot of functionality, and so how do you think about A/B testing within that that framework?

Vamsi Kodam (05:43) the way I look at it a lot of literature about A/B testing is in an e-commerce framework where you're shopping on a website, you want to check the conversion or et cetera. But even in a like a B2B kind of a context like the the space that I I operate in, there are decisions that need to be

Vamsi Kodam (06:01) made and there's always okay I'm doing it this way and there's other way to do it and which one is better. for example the one experiment that I ran fairly early in my career that had that left a very big impression on me is around how airline was using AI this is like 15 years ago

Vamsi Kodam (06:20) I'm talking about not the LLM AI, but

Ashley Stirrup (06:22) Sure.

Vamsi Kodam (06:23) the the the time series and a lot of the machine learning models.

Ashley Stirrup (06:27) Right.

Vamsi Kodam (06:29) so they use this technology to come up with the decisions, right?

Vamsi Kodam (06:33) And the results that came out of it were very revealing and so in the end the result was that this particular A/B test like revealed that net net overall, it wasn't bad. It was more or less the same, both using the human intervention or without human

Ashley Stirrup (06:50) Right.

Vamsi Kodam (06:51) intervention, like the revenues

Vamsi Kodam (06:52) were more or less the same. So the revenues did not

Ashley Stirrup (06:54) Yeah.

Vamsi Kodam (06:55) differ much. But what was actually more revealing was this is at a network level like so that means you have like different points of sale like different when you slice that data there was a lot to uncover and we split it into groups like where there were parts of the network where the human intervention helped make more revenue.

Ashley Stirrup (07:18) Hm.

Vamsi Kodam (07:19) And then there was parts of the network where human intervention or without human intervention did not make much of a difference. And there were parts of the network where human intervention actually made the revenues worse. Like you didn't you left the system untouched, it would have actually done a better job. And this

Ashley Stirrup (07:34) Right.

Vamsi Kodam (07:35) so it was not like a win or lose, right? It's actually a learning of how you interact.

Vamsi Kodam (07:41) and that led to

Vamsi Kodam (07:44) changes in the practice, where okay, hey, these are the kind of scenarios where the human intervention helped, and these are the kind of scenarios where human intervention did not help. it provided a guideline to the human intervention. So take this,

Ashley Stirrup (07:55) Yeah.

Vamsi Kodam (07:56) we implemented it, and then like we did another round of A/B testing.

Vamsi Kodam (08:00) this had a kind of a formative impression on me, like in terms of experimentation and openness to A/B testing, because I had a certain intuition that this is how it should work. And

Ashley Stirrup (08:10) Yeah.

Vamsi Kodam (08:11) but the results were kind of counterintuitive and then there were

Vamsi Kodam (08:15) learnings. There's nothing like a failed experiment, I would say. There is there's

Ashley Stirrup (08:18) Yeah.

Vamsi Kodam (08:18) always a learning and you don't do something that is actually a win as well, right? Because you you don't go on the go down that path.

Ashley Stirrup (08:25) Right.

Vamsi Kodam (08:25) you improve what you're doing. when there is a best practice or a intuition, I would not take it on faith but

Vamsi Kodam (08:33) l I've learned to question and kind of think about it from first principles and collect data to back

Ashley Stirrup (08:38) Yeah.

Vamsi Kodam (08:38) the decisions

Ashley Stirrup (08:51) Yeah. Yeah. It's funny 'cause so often, you'll be so sure that a feature's gonna win and then it it won't. and sometimes it might be that's a bad idea. And sometimes

Vamsi Kodam (09:02) Mm-hmm.

Ashley Stirrup (09:03) it might be, well, it was a good idea, but we didn't quite get the implementation right. And the chances of getting it right on the first time are actually pretty low. So it's not surprising

Vamsi Kodam (09:12) Pretty low.

Ashley Stirrup (09:13) that you have a a low win rate.

Ashley Stirrup (09:14) I think that creating that humility is one of the most powerful things. That once you get to that point of humility, then you realize, okay, I I do need to test everything. And that opens up a world of learning that otherwise people just totally miss. So this example sounds pretty interesting. was there like a pattern where the humans were always raising prices or lowering prices, or they were just adjusting them.

Ashley Stirrup (09:36) more intuition based than formulaically, yeah?

Vamsi Kodam (09:39) yes, I think the biggest return was like around a human intervention that was automated. they had some kind of an event for which they assumed that the prices would be higher and they went and made some adjustments to increase the prices.

Vamsi Kodam (09:52) but they forgot to remove that intervention and that kept applying and it went unnoticed.

Ashley Stirrup (09:55) Hm. Right.

Vamsi Kodam (09:57) I would say like generally the humans have an intuition of like how this will work,

Vamsi Kodam (10:01) it's not always like backed in data.

Ashley Stirrup (10:03) Yeah.

Vamsi Kodam (10:03) connecting that intuition to the data can help. And j just being open to taking that data feedback.

Vamsi Kodam (10:11) And then seeing what's working and what's not working and then putting that into whatever that we are doing. It's a a new feature that we're implementing or a new workflow or a a process. it's always beneficial.

Ashley Stirrup (10:21) Yeah, long ago I worked at a supply chain startup and we helped people implement Kanban, just a very simple

Vamsi Kodam (10:29) yeah.

Ashley Stirrup (10:29) inventory replenishment process. and what we were able to show over and over again through statistics is that humans, you know, thought they knew better and they would order more, but they would do it sporadically and it just not do it in a consistent way. And that

Ashley Stirrup (10:45) that lack of consistency actually introduced a lot of cost in the system, both stockouts and too much inventory. And so even though it felt good to the humans, it it often wasn't. So when you started telling that story, I was kind of expecting the the system to always win. But sounds like in your case the humans were able to identify pockets of opportunity, but other times there were, you know, benefits from having the systematic approach.

Vamsi Kodam (11:10) Yes. Yeah. Yeah. And

Ashley Stirrup (11:11) Yeah.

Vamsi Kodam (11:11) i i it's always like I I would say it's a 80/20 or a 90/10 where a well designed system should take care of most of things, but there are always this edge cases, the guardrails,

Ashley Stirrup (11:20) Yes.

Vamsi Kodam (11:21) that the systems tend to be solved on, but over time systems also got better.

Ashley Stirrup (11:25) That's right. But a good opportunity for continuous improvement. and so at FLYR, how often are you running A/B tests and how often are you choosing to do other approaches in order to kind of understand the impact of features,

Vamsi Kodam (11:38) yeah, so at FLYR we implemented reinforcement learning for a pricing problem. this is more advanced experimentation multi-arm bandit experimentation where there was no data like for this particular pricing problem and so what we had to do was

Vamsi Kodam (11:57) create multiple variants and then put traffic through those multiple alternates and then learn which ones are winning and losing and then use the data to then route more traffic to the ones that are winning and compared

Ashley Stirrup (12:13) Yeah.

Vamsi Kodam (12:13) to the biggest difference I would say compared to a traditional A/B test is how fast it is to

Vamsi Kodam (12:20) learn from the data. For example in a traditional A/B test you design the experiment, you run the experiment, you wait for the results, you do the analysis and then you declare victory or not and then take the

Ashley Stirrup (12:31) Yes.

Vamsi Kodam (12:32) next steps. And there's a that window that it takes and and with a multi arm bandit the advantage is that you you don't have to wait for the experiment to conclude and based on the data like it can

Vamsi Kodam (12:43) start routing like more traffic to the winning experiment. And at the same time you're not abandoning other arms, but you are still testing and if they are if there is a change in behavior, et cetera, it'll adapt that trend and route more traffic to that. this is something that we implemented for a pricing problem at FLYR and

Ashley Stirrup (13:01) Yeah.

Vamsi Kodam (13:01) the the airline

Vamsi Kodam (13:03) has seen l double digit revenue improvement

Ashley Stirrup (13:06) That's great. Super interesting. So, how how does it work at FLYR? I mean, do you have like a centralized data science team? Is each product manager doing their own A/B tests? How how does that work?

Vamsi Kodam (13:18) Yeah. So airlines are our customers so we are working with our airline partners to do this and then we provide the product platform that helps airlines do this. So we are providing this to airlines as a product offering.

Vamsi Kodam (13:32) And

Ashley Stirrup (13:32) Got it. So

Ashley Stirrup (13:34) you've actually built it into your product then. Yeah,

Vamsi Kodam (13:36) Yes.

Ashley Stirrup (13:37) that's super interesting. 'Cause then all of a sudden you have a ton of traffic, right? if it's one thing to be optimizing for airline employees, but it sounds like you're also helping the airlines do A/B tests with their customers.

Vamsi Kodam (13:51) yes. so we have multiple offerings, some of our products are sitting in the background, right? all the consumer traffic for the airline directly hits like our products so that it's thousands of requests per second and okay, and

Ashley Stirrup (14:03) Yeah.

Vamsi Kodam (14:03) these are all responded to

Vamsi Kodam (14:06) in in low late latencies, we're talking twenty, thirty milliseconds,

Ashley Stirrup (14:09) Yeah.

Ashley Stirrup (14:10) Yeah. Super interesting business opportunity. obviously like you said, the traditional airline stack is pretty old these days. but I would imagine unseating those is not a not a simple thing either. when it comes to d designing experiments,

Ashley Stirrup (14:27) Are there things that you try to do to make sure that you're learning as much as possible from them?

Vamsi Kodam (14:32) yes.

Vamsi Kodam (14:33) The one thing that has benefited me the most when it comes to running experiments or picking any new problem and trying to make progress on it is I look at like a l a lot of existing literature and not academic ones, but there's a lot of good practice literature as well. Like a lot of practitioners publish

Vamsi Kodam (14:51) results and the findings, etc. I whenever I start a new problem, be it experimentation or even a complete new problem surface, I do a lot of discovery we can't be the the first ones to be facing this problem

Vamsi Kodam (15:04) which other industries or which are the companies may have faced this problem or solved this problem or like are likely to have faced this problem and solved this problem. When you explore

Ashley Stirrup (15:12) Yeah.

Vamsi Kodam (15:12) in that dimension, that led me to a lot of good practice papers and sometimes like the technical blogs that gave me a lot of good insight

Vamsi Kodam (15:20) and then instead of starting from zero, when I do an experiment, I have a built a intuition around okay, this is what people have done, is how it would it would translate to the experiment that I'm trying to run and kind of go from zero to one.

Vamsi Kodam (15:33) faster, you

Ashley Stirrup (15:33) Yeah.

Vamsi Kodam (15:34) you can build on a lot of learning that's out there. And that's something that has helped me. for example like when I l ran my first experiment, I didn't do that. So I actually kind of did that from scratch. But these days when I look at experiments

Vamsi Kodam (15:48) I look at the technical blocks, for example, like recently I think DoorDash I saw they they did like some ex experimentation with like multi arm bandit, how they rolled it out, etcetera. So I see a lot of learning from them and and there's a lot to pick

Ashley Stirrup (16:02) Yeah.

Vamsi Kodam (16:03) from what others are doing and you don't have to do everything on your own.

Ashley Stirrup (16:07) I think it's a really great call

Vamsi Kodam (16:08) This.

Ashley Stirrup (16:08) out. I totally agree, especially for people in lower traffic scenarios to go look at the people who have a lot of traffic and what kind of tests have they run and how have they worked. at the same time you as a a product manager,

Ashley Stirrup (16:21) You're probably juggling a lot of balls and so it can't be easy to carve out the time to go look for, you know, what other research has happened that's kinda similar to what I'm considering.

Vamsi Kodam (16:32) y yes, it used to be harder, but like now I actually have AA bots that are set up like to monitor

Vamsi Kodam (16:38) the content and it i it covers actually like a lot of space and I get like some interesting summaries of the

Ashley Stirrup (16:43) That's a great call out. Yeah.

Ashley Stirrup (16:45) Yeah. Super interesting. and so as you kind of look forward for FLYR, what do you see as the the big opportunities in terms of where you'd like to see experimentation evolve there?

Vamsi Kodam (16:56) Yes. so we are building the modern commerce platform for airlines so that they can provide this personalized retailing experience. And one of the opportunities that I see here is generally with like airlines, a lot of focus historically for airlines has been around the price, lot of

Vamsi Kodam (17:14) rigor in terms of like what price you're selling it at. And but increasingly that is moving towards it's not just like price, but it's also like about the product. For example, let's say when you go to an airline website and you shop for like okay, you want to go from A to B, typically you see the standard three or four offerings and they are relatively the same regardless of your context.

Vamsi Kodam (17:36) and airlines are known to change prices, like dynamically, but going beyond that, right? there's a opportunity to experiment on the product. Let's say so price is one of the variables, but let's say like you're a leisure traveler, you're going on a vacation with your family and you're booking like two months in advance, for a week, etc.

Vamsi Kodam (17:55) in such scenario in addition to a specific bundle or a product offering that'll likely make more sense for you versus let's say you're going on a a business trip like next week for like two days w where you're booking at much closer to departure and you want shorter length of stay, et cetera. So your preferences will be different. And I see a l

Vamsi Kodam (18:16) lot of opportunity going beyond the price, right, experiment on the product preferences like kind of how these resonate with the the travelers. I I see a lot of opportunity there. And so we've built a lot of foundation for the commerce platform so we lay in the groundwork and one of the opportunities that I see in the next one to two years is

Vamsi Kodam (18:36) building this product side of the experimentation get the learnings out of it and then iterate to

Ashley Stirrup (18:41) Yeah.

Vamsi Kodam (18:42) use that product variable

Ashley Stirrup (18:44) Yeah.

Ashley Stirrup (18:45) Are you are you building any LLM powered features into your platform?

Vamsi Kodam (18:49) not actively so not directly the product at this time. we are providing this API products like to our customers that they're using they they doing the consumer experience and then they also have the LLMs on the airline side. So we're not doing this. But

Ashley Stirrup (19:02) Got it. Got it.

Vamsi Kodam (19:06) in the last six to twelve months, the the biggest benefit as a product person like from LLM or the AI technology is around the speed of development. So some of the things that used to take, let's say, a a year or two and are done

Vamsi Kodam (19:19) in like a quarter

Vamsi Kodam (19:20) Things are moving fast and and so that gave us like opportunities to address a lot of tech debt more proactively and also pick on a lot of features that could have taken much longer. So we were able to move faster

Ashley Stirrup (19:33) Yeah.

Ashley Stirrup (19:34) Yeah, no question. There's a lot of opportunity to move faster with AI. we hear from a lot of our customers at Growthbook that they wanna make sure that they're wrapping every new feature in a feature flag so that they can ship fast but still have those kind of auto rollback features that you get with feature flagging so that

Ashley Stirrup (19:53) they've got that safety net 'cause, obviously the faster you go, the more opportunity there is for something to break. And for sure, something like the airline industry, that's not something you you want.

Ashley Stirrup (20:03) Well, Vamsi thank you so much for joining

Vamsi Kodam (20:04) Yeah.

Ashley Stirrup (20:05) today's episode. I think it covered a r really interesting industry and sounds like FLYR has a bright future ahead of it. I I look forward to keeping an eye on the business and seeing how it it does going forward.

Vamsi Kodam (20:18) Thank you, Ashley. it is great talking about FLYR and my experience with experimentation.

Ashley Stirrup (20:24) Thank you so much.

No items found.

Subscribe to the podcast

Takeaways from this conversation

As AI accelerates development, feature flags help teams release more safely.

S2 | E13

Beyond dynamic pricing, airlines can test which products and bundles best match traveler needs.

S2 | E13

Multi-armed bandits can shift traffic toward better pricing variants while continuing to learn.

S2 | E13

Test human intervention against automated decisions, then define where overrides add value.

S2 | E13

A tied overall result can reveal meaningful gains when you analyze individual markets or customer segments.

S2 | E13

Top takeaways from other favorite conversations

All Takeaways

The most valuable North Star metric is the one you can't measure yet, long-term client value, and causal-inference modeling helps predict it from short-term behavior.

Go to S1 | E24
Theme
Future of Testing
Role
Data Scientist
Industry
Retail
Featured
false

Keep the experimentation analyst team independent from the product owner to reduce cherry-picking.

Go to S2 | E12
Theme
Culture
Role
Product
Industry
Retail
Featured
false

Start with the advertisers and product categories current technology can support, then expand with evidence from real users.

Go to S2 | E11
Theme
Scale
Role
Product
Industry
Retail
Featured
false

Build hypotheses around user psychology, not just KPI movement

Go to S1 | E9
Theme
A/B Testing
Role
Product
Industry
Media & Gaming
Featured
false

Build an AI ecosystem with clear purposes (productivity, engineering, consumer) and a steering committee to avoid duplication.

Go to S1 | E5
Theme
Culture
Role
Exec
Industry
Financial Services
Featured
false

A test that wins on the primary metric can still be a loss. Expedia's sign-up takeover raised conversion and damaged return rate and engagement at the same time.

Go to S1 | E44
Theme
A/B Testing
Role
Data Scientist
Industry
Business Tech
Featured
false

One or two big wins a quarter is a healthy hit rate when you run 150–200 experiments a year.

Go to S1 | E17
Theme
Scale
Role
Data Scientist
Industry
Business Tech
Featured
false

The three-click rule is conditional. Clicks only hurt when they're empty; a click that narrows thousands of options to dozens is a feature, not a cost.

Go to S1 | E16
Theme
A/B Testing
Role
Exec
Industry
Retail
Featured
false

Testing gets watered down when "let's try something" replaces a control group; a little pre-planning gets far more out of every experiment.

Go to S1 | E32
Theme
Culture
Role
Data Scientist
Industry
Financial Services
Featured
false
The experimentation edge podcast logo with a picture of host Ashley Stirrup