Jul 28, 2023
In May, Oscilar co-founders Neha Narkhede and Sachin Kulkarni were the guests of the PitchIt podcast by Fintech Nexus, hosted by Todd Anderson.
In this conversation, our founders and Todd Anderson (Chief Content Officer of Fintech Nexus), discuss the founding story behind Oscilar and the importance of no-code and AI-powered risk solutions in the rapidly evolving fintech landscape.
They also highlight the significance of diverse engineering talent, the growing challenge of fighting fraud, being married as co-founders, the journey of bootstrapping the company, and much more.
Throughout the conversation, Neha and Sachin showcase the rich expertise they bring to Oscilar, drawing from their previous roles at Confluent, Facebook, and LinkedIn, to underscore the critical role of real-time data and machine learning in tackling complex problems like fraud detection and risk decisioning.
You can listen to the podcast here.
On today's episode, I'm joined by Oscilar co-founders Neha Narkhede and Sachin Kulkarni. Neha, Sachin, and I talk about the founding story behind Oscilar, the importance of no code, the efficiency of out-of-the-box solutions, which right now that's what Oscilar is offering, out-of-the-box solutions, the benefits of having engineered talent from the Ubers and Apples of the world: it's really diverse in terms of the talent that they have at Oscilar, the rising importance of kind of where fraud fits and fighting fraud within organizations, being married as co-founders and kind of the uniqueness that comes with that, raising capital and much, much more.
One quick PSA:
If you want to reach the best Fintech community around, we have all the tools you need to shape your message and reach new clients. You can sponsor an episode, sponsor a webinar, buy ad space. We have it all here at Fintech Nexus, so please reach out anytime.
Happy to chat about any of our digital products. [email protected] and my email address will be in the show notes.
Without further ado, I present Oscilar’s co-founders Neha Narkhede and Sachin Kulkarni. I hope you enjoy the show. Good morning, Neha and Sachin.
Welcome to the podcast. How are you both doing?
Doing great. Great. Thank you so much for having us.
Of course, so I like to kick off the episode if we can just get a little sense of your story.
So maybe Neha, I'll start with you. And then Sachin, I'll go to you next. Tell us a little bit about your journey, kind of where you've been professionally until now, and kind of what brought you ultimately to the point of wanting to start a company together.
Yeah, I'm co-founder of Oscilar and also co-founder of Confluent, which is a public tech company. Previously I led product and technology at Confluent, where I was privy to a lot of applications of event streaming and real-time data. Before that, I led data streams charter at LinkedIn, where me and my co-founder started Apache Kafka, which eventually led to Confluent.
And the reason for starting Oscilar in probably a really short way is that I saw a lot of use cases on real-time data, and the fastest growing and the most interesting one was fraud detection and risk decisioning. And that is what led to Oscilar.
Hi! So before Oscilar, I used to be at Facebook for more than a decade. Actually, eleven and a half years-ish where I was the Senior Director of Engineering responsible for building out Facebook's private cloud. Also built the back end for Facebook live Facebook videos, Instagram live, Instagram videos, and the back end for Facebook Messenger.
And so along the way, we improved the trust and safety of the Facebook platform overall by building hyper-scale real-time data processing systems that detect various types of fraud, as well as things like video integrity, people trying to video copyright and so on, like those types of problems as well. And I think that there was a fun fact there where my organization, the video infrastructure organization, eventually won an Emmy in the technical category, of course. And so there was a running joke among my friends where they would, of course, make fun of me and call me an Emmy Award-winning director.
Not a bad thing to be called.
So I always ask the question to my guests, but what drives you, and ultimately what drove you to want to start your own business? I mean, I hear stories like my parents were entrepreneurs, and it was kind of ingrained to me.
Neha, you mentioned kind of finding a problem through your other journeys in terms of the businesses that you were at that you wanted to start Oslor in particular. But where does the drive come from when it comes to being an entrepreneur?
For me, it was really the complex solution that risk decisioning really is today for a really specific and growing problem, which is the rise of online fraud and risk. So the context behind that and the way we arrived at this is over the past decade, as companies became more digital, there's an explosive rise in online transactions, and that's further accelerated by the pandemic.
So today, many of us expect to handle the most consequential actions of our lives, whether it's applying for a home loan or transferring money through a peer-to-peer wallet. We expect all this to happen online, instantaneously, and also securely. And the sharp increase in online transactions has actually resulted in a dramatic increase in risks of all manner.
It's not just fraud risk, it's credit risk, insurance risks, and so on. And the reality is that most companies risk technology that has not kept pace with this proliferation of transactions. And it's not that these costs are substantial.
When I saw the cost, credit, and fraud risk now cause consumers more than 8.8 billion annually. And personally, I saw this trend unfold in real-time at Confluent when I witnessed a large variety of applications using event streaming data and this being one of the major ones where companies had to string together lots and lots of distributed systems along with machine learning pipelines in order to solve this problem.
And even then they had this problem of dependence on engineering not being able to solve this quickly enough. And before we jumped in, I talked to hundreds of fraud and risk leaders to just see what their challenges were. And it turns out that they face similar challenges: providing autonomy to risk teams, ensuring access to the most relevant data at the right time, developing relevant machine learning models.
And therefore the complexity of this solution was really intriguing to me. And Sachin can say more, but I wasn't alone in wanting to build a new solution here.
Yeah, I think what we realized is we were at some of the top tech companies with great talent and we still found that the existing systems were cumbersome and in many cases ineffective in dealing with fraud.
And it was not just the companies we were at. We also talked to people who had worked on these systems, let's say Twitter and some other companies, and got the same sense that this is we were actually confounded by how bad the situation was, and it helped us realize that this is actually not just in these companies. If you go outside the top tech companies, the problem actually gets worse because they don't necessarily have access to the talent that these companies do.
And so we were convinced that this problem needs to be solved. There was also a personal anecdote that said absolutely, there’s a lot more, right? Where I think one of my older uncles, he is in his 70s, he got a phone call one day saying hey, there has been some bad activity on our account so share the code that we sent to your estimates so we can reverse that activity.
Of course, this was a scam, but I mean him being in his 70s is not really up to speed on everything that is happening in different kinds of scams. So he gave them that number and they of course drained his account right away. And that got us thinking right.
This is a real problem for real people. There's not just sort of at the fringes anymore. People are losing their life savings.
And I don't think it is reasonable to expect that every one of these people get fully educated on how to stop these things. That's not going to happen. We need systems in place.
We need businesses to be able to do this for their users, protect them from this type of activity. And so I think the personal as well as the professional sort of those words collided, if you will, in helping us figure out, yes, this is what we need to build and solve this problem.
So as we jump further into our conversation a bit more about the solution that you're building. Where did the name Oscilar come from?
Yeah, online risk is never really static. Like all the leaders we talk to and our experiences as well pointed to the fact that companies have to pick somewhere on the spectrum of ease of use and catching every single fraudulent activity they want to have. And it's usually a spectrum.
You have to pick different points at a different stage of your company and depending on what you are seeing, we would like people not to have to make that choice. Like ease of use should be there while you catch all the fraud, but not possible in every case. Sort of that oscillation.
It was an interesting spectrum of the balance. Very okay, Oscilar is similar enough to Oscillation.
So tell us a bit more about the company. Is there kind of a perfect type of client that you serve and kind of where are you guys today in terms of your journey? And Neha, tell us a little bit more about what you're building at Oscilar.
Yeah, so over the past two years, we've been working in stealth alongside a world-class team of engineers to build this first-of-a-kind real-time AI-powered risk platform that dramatically improves the quality and risk speed of risk assessment. We publicly announced the company last month, and the reaction has been very positive.
And we're looking forward to growing the company significantly in the coming months and years. And where we are on the customer development journey is we're working with dozens of fintech customers. For instance, we partner with leading fintech companies like Slope and Super for applying our technology to help them secure and reduce the risk of their online transactions, both B2B as well as B2C, and our products currently are twofold:
We offer a fraud detection product as well as a credit underwriting product, and usually customers start with one product and very seamlessly upgrade to the other one.
So really we're looking forward to growing the reach of our products across industry sectors with the chance of building a really iconic company that can make the internet safer.
And so when doing research for this episode, looking at your backgrounds and a lot of the information you have on your website, obviously you make it very apparent about the no-code portion of this solution.
So why no code? And then first tell us a little bit more about the benefits, and then how much of a learning curve is there amongst financial service executives? Some of the clients that you mentioned with no code, from my perspective, I'm starting to hear about it a lot more in fintech and financial services. But tell us a little bit more about that decision and kind of how clients are adapting to what you're building.
Yeah, great question.
So I think as part of research, one thing we realized is the risk teams, the risk operations teams or analysts, and so on, they have a lot of domain knowledge about the business, the types of fraud they want to catch, and so on, right? But may not necessarily be the ones that write the code. And so we had this separation of responsibilities where there was the risk team that knows the business and then there is the engineering team that builds the system. And we realized that that separation was actually causing several problems.
The risk teams were not autonomous. They had to constantly depend on engineers to go build things or make small tweaks, and that was slowing them down and making their fraud detection efforts much less effective. I think that's where no code comes in, where you don't have to depend on engineers to make small changes.
The engineers can focus on the core product, while the risk operations team can try out many things, experiment very quickly, and apply all their business knowledge to solving the problem without being blocked. And so that autonomy was really important. And so we have built the system in a way that teams can get started in minutes.
It's intuitive, you don't have to know anything else. There is no new DSL that they need to learn. As long as they know SQL, they are good to go.
And essentially we have seen most analyst teams know SQL already, so they can get going. And it's very consistent with their experiences and what they need to know. And so getting that risk signal, starting from okay, there is a fraud pattern emerging to finding a solution to it, deploying it, then seeing that fraud go down or get detected quickly, that should happen in minutes.
And with that as the goal, no code was the way to go. We could not have done that without no code.
You also make a point to say the solution is out of the box.
How much faster is that for clients when they're saying, all right, I have all these fraud potential options, I have all these credit options, but I can't afford to do a twelve-month proof of concept or 18 months? And all of a sudden now we're starting to get going. We need to get going today, I need to protect my customers today, and I need to continue bringing in new customers.
So how important was that to focus on out of the box? And how much faster does it get to market versus maybe the traditional POC-type way?
Yeah, absolutely. So I think out-of-the-box solutions are incredibly important because of the time to market.
I think you hit the nail on the head. There are so many risk vectors at play that people have to deal with, expecting that they put all of those together in a very short time frame. Doesn't happen if you go the POC approach.
It does take, like, several months to get everything together, and so on. And so the out-of-the-box play is actually really important. It allows them to go to market quickly, but also scale existing ones or add new products, new capabilities on top of existing products.
And so having ML models out of the box that we offer gives them a much shorter runway in getting to market. That doesn't preclude them from doing sort of the detailed tuning or having their own machine learning models and so on. So it's not one or the other.
So people use one to get started out of the box to get started. And then over time, as you learn more, they can add their own custom features on the same platform. They don't have to learn a new platform or do things differently.
So I think that combination is very powerful.
You kind of hit on my next question, which was, are most clients or those that you're talking to looking for one of your solutions? Kind of the whole suite of solutions? How are they looking at the various things that you provide right now?
Yes, so Oscilar is designed to be a modular platform.
And so it's not all or nothing, right? And so most customers start with one solution. And we have several. We have solutions around credit underwriting. We have solutions for account takeover, synthetic ID, transaction monitoring, and so on. And so most customers start with one and then over time expand into another. That makes the whole process much simpler for them.
They're solving one problem, they see success there. This works really well. Great.
Now let's start adding more. So we have seen people start on transaction monitoring over time, start using the solution for credit underwriting, and vice versa. So I think both are possible.
That upgrade is seamless. You don't have to do sort of a whole separate thing. And one of the reasons why it is seamless is we don't impose a specific data schema.
We found that customers were spending tons of time doing data mapping between their internal data format to the format that was imposed by a vendor. And that was a no-go. Right? Because that's going to take weeks to months.
And that's not the time that people have or should be spending on this. And so we don't impose a schema, which means they're spending zero time on doing any mapping. Whatever schema they have, send that data in.
We will figure it out in sort of almost automatic fashion.
We've talked a little bit about your backgrounds. You also have people in the team that have worked for Apple and Uber and LinkedIn.
I mean, how important is it when designing a solution or solutions like you are to have those different perspectives come to the table and pair that with your backgrounds because you have so much experience, whether it be at Apple or Uber, that gives you the perspectives that the customer might ultimately have.
Yeah, absolutely. So I think the diversity of perspective is actually incredibly important.
If you look at our backgrounds and where the different people that we have hired and our founding team is from. There are social networks, there are sort of the marketplace like Uber, Apple, one of the top tech companies in the world, very different experiences, and those help us design a solution that will work in a broad set of cases. And so I think that diversity was important, we were very intentional about hiring people and getting the founding team with a lot of this experience.
And this came up in our conversations with the 100-plus risk leaders that we talked to before starting the company. It was clear that the challenges are similar, but there are enough differences as well that need to be accounted for in the solution.
And so I think that's where the diversity of the engineering team, to begin with, was actually quite good.
And these days, you have Uber that is becoming a financial services company, Apple's becoming a financial services company. I mean, it's starting to mix all across the different industries that these types of solutions. I'm sure as a founder yourself, you're looking, all right, fintech is here, but Apple is becoming a fintech and others are becoming a fintech.
The universe is expanding.
You're exactly right. That's what we see.
That's the trend. And that's one of the reasons why we're so excited about bringing this kind of solution into the market.
Yeah. On the fraud piece, you obviously touched on it earlier in our conversation with the onset of COVID and we went to so much digital transactions or more digital transactions that we've ever seen before in a pretty short time frame. Do you think fraud is finally getting the attention it deserves across fintech and kind of financial services, that it's kind of risen to almost the top of the stack? And do you think with that fraudsters share openly in public? Can the good actors or the fighters do better at maybe sharing with it? Not necessarily in a public space with a lot of the stuff they deal with, but in some ways that can help each other, because ultimately, if J. P. Morgan's hit, it doesn't help Bank of America, even if they're competitors.
Yeah, exactly. If you study the last five years and the trends, there certainly has been a growing attention due to the massive problem of fraud.
Right? And if you look at the banks in the past, they had formed a consortium and they have that where they band together and from a payment rail perspective have something going there. I think we need a similar open consortium of all fraud vendors where everyone can come together and foster this data sharing, which makes everyone's products better as compared to when they were operating in their own landscape. But what's happening right now, at least what I see is there are some startups that are trying to form their own consortiums and trying to keep it closed to their own customers, and that just limits the data that their customers can get from just themselves.
And I'm really looking forward to a consortium which is more independent, that can bring together all the vendors and then therefore be a much more powerful tool in our toolkit to fight fraud.
You've mentioned you've been in stealth for a couple of years. What’s, if not the biggest, what are some of the lessons that you've learned from when you started to when you just launched recently, had all that time in stealth? What are the things that you've learned, maybe about yourself being a founder or about the company that you can share with others that might listen to the episode?
Well, about myself, what I've learned is to not compare Oscilar's journey with my previous company, Confluent, where we saw a large amount of success in a very short amount of time. And if there's something I would change about our Oscilar journey, it is that we should have started the company much earlier than we actually did.
Some amount of research is very useful, but starting it earlier would have been even more useful.
And I think in some sense the timing was really right. Neha being the co-creator of Kafka and the co-founder of Confluent, it was clear that real time and real-time AI is in our DNA.
The technology has finally caught up to what the world needs to catch fraud very quickly. And so I think that came through together very nicely. And so, yeah, maybe a little bit before would have been fine, but I think the timing is just right for the technology to catch on.
Tell us more about how big is the team today? You mentioned top engineering talent. What's the makeup of the team? Is it 90% engineers? Tell us a little bit more about those that are around you, helping you to build Oscilar.
So we have a fantastic team.
We have about 30 people, mostly engineers with decades of experience from Facebook, Now, Meta, Google, Uber, Confluent, and so on, who have deep expertise in building highly scalable real-time AI systems. And so I think that's the expertise we have been looking for. We also have analysts and data scientists who have dealt with fraud and machine learning all their careers effectively.
And so they also bring an interesting and different perspective on how the system will get used in the real-time, sort of perspective on what we need to build, not just the engineering way of building things, but also a user view on what needs to happen. So I think that combination has been great, and I think in the post-COVID world where being remote-first or virtual first is the norm, of course, our team is spread out through US and Europe. I think that sort of diversity of ideas is also helpful because even though there are many similarities, there are also differences in how things get used in the fraud patterns across Europe and the US.
Yeah. I see that you've raised some outside capital from investors. How was fundraising maybe compared to previous ventures, at least for you Neha, what did you learn that maybe you didn't know about Oscilar through those investor pitches and talking to investors that maybe we should do this or potentially implement that along the way? Do you learn stuff like that in pitches? Tell us a little bit more about how fundraising went.
Yeah, so as part of our public launch, we announced that we are self-funding the company with $20 million, opting to not take outside funding so that we could quickly build and scale the company. We are fortunate to have received strong interest from investors, and we look forward to exploring fundraising opportunities at the right time in the future to bring in the right advisors to the company.
So we have just a few minutes left.
I like to end a little bit lighter than we started. And so a couple of fun questions to end. But first, I know that you two are married. How is it working with your spouse? Were there any reservations? Ultimately, when you're thinking of starting a company or working at the same company together, does anything come into play like, hey, maybe we shouldn't do this? Just a curiosity from my standpoint.
Neha, you want me to take this?
So we actually talked to several co-founders who were couples themselves to learn from their experience on what worked, what did not work.
And one thing they said consistently is, have a clean separation of responsibilities. If you have that, it increases chances of success dramatically. If you don't have that, you will fail.
You have a high chance of failure if you don't have a clear separation. And so before we started, we were like, okay, this makes sense. Let's have a clean separation that way. This was a clear lesson.
And so we did that. So Neha focuses a lot on the go to market, the product aspects of the company, and so on, and I focus on the engineering aspects, building out the product, and that also plays to our interests and strengths. And so it has worked out really well. I would do this again.
So onto those questions. Do you have a favorite book or the last book that you read?
Well, I have too many favorite books to name one, but the last book I read, which really fascinated me, was Outlived by Dr. Peter Attia.
Sachin, any favorites on the book side?
I think on the book side, I'm spending most of my time reading about raising children.
Yes, children… I have two. So it's an adventure that never ends, that's for sure. Obviously, as founders, you need to be fresh and present for your team. Do you do anything specific to unwind, take your mind off work and not burn yourselves out?
I love to unwind by playing with our son.
It's really fun to see the world through his eyes and really sort of take the weight off your shoulders.
I like to play soccer, so every Sunday soccer time gives me good exercise and then I feel sort of, okay, this is good, Feel better, meet some friends along the way and so on. And then of course, playing with my son is the highlight of the week, I would say.
You mentioned soccer.
I'm assuming that's your favorite sport. Do you have any other sports that you play or sports teams that you might prefer?
I don't play American football, but I root for the Steelers. I went to Carnegie Melon in Pittsburgh.
They won the Super Bowl that year. That was the first year in the US. Been following them since.
Fantastic. Not doing so great recently. I hope they get back to their many ways soon.
Yeah, we'll see. Favorite vacation spot?
For me, it's the Italian Dolomites.
I like to go to India.
It's a great combination, meeting family. There is a lot of interesting food. I have a list of 20 places that I want to eat at every time I go.
Different places to visit. The culture is sort of colorful, fun. That's my favorite place to go.
And then the final question, what's your biggest inspiration? What inspires you to kind of keep going, get at it every day? What's inspirational to you?
For me, it's my dad who inspired me to be an independent thinker and an entrepreneur. And I strive to do this and continue to be successful at it to make him proud from wherever he is.
For me, my parents as well, it was an interesting combination that they also have a business together that they run and operate and very different personalities too.
My mom is “go get things done” quickly and so on. And my dad is more of a thinker and so that combination worked out quite well for them. And Neha and I are not that different.
We are similar in many ways, but yeah, by learning from my parents how they operated this and how they went about starting a business, it was not easy for a woman to start a business in those days. This was several decades ago, so, yeah, I think I would say my parents would be my inspiration.
Well, Neha and Sachin, I greatly appreciate you spending some time with me, and thank you for coming on the show. If someone wanted to reach out and start a conversation or wanted to talk to you, how can they do that? How can they reach out to you or how can they reach out to Oscilar?
One way to reach us is just our about us contact page. The other way to reach me is LinkedIn.
Okay, well, again, appreciate you coming on the show. Continued success to you and the team, and hopefully we'll get you back sometime in the future.
Thank you for having us. It was a lot of fun.
Of course. Thank you.
Thank you, Todd.
Ready to give autonomy to your risk team?
As you can see, at Oscilar we are building a risk operating system that allows risk teams to take full control over their operations - without the need to rely on expensive engineering resources.
After talking to hundreds of industry professionals, we have realized that despite the different kinds of fraud challenges in various verticals, there is still a lot to improve regarding real-time detection & prevention capabilities that rely on 100s of data providers. That’s why our engineering team has experience with companies where these problems require solutions at scale.
If you are curious about what the platform can do in action, consider booking a demo with us below.