Employee Spotlight
Greg Kuhlmann — Co-Founder & CEO
Sumatra’s Greg Kuhlmann shares how he got his ML street cred from Apple, how he transitioned from data scientist to startup CEO and co-founder, and what happens when you take your foot off the gas (hint: you roll backward.)
Jun 16, 2023
We sat down with Sumatra’s Greg Kuhlmann to learn how he got his ML street cred from Apple, how he transitioned from data scientist to startup CEO and co-founder, and what happens when you take your foot off the gas (hint: you roll backward.)
How did you end up working on Fraud at Apple?
As I finished my Ph.D., a grad school friend reached out about an opening on his team at Apple, using machine learning to fight fraud on the App Store. I knew nothing about fraud — I had never owned an Apple product. But I guess I had enough ML cred to pass the interview.
Over the next eight years, I worked with an extraordinarily talented and humble group of people on projects with dizzying levels of impact. It’s harder to do ML at a company that genuinely cares about privacy. Fighting fraud at a company obsessed with user experience is even harder. For those reasons, I’m confident that Apple has the best anti-fraud data science team in the world, bar none. I feel grateful to have been a part of it.
How has the transition from a data scientist at Apple to Founder of a startup gone? Any lessons learned?
Large organizations, at their best, are these indomitable forces that move projects forward, regardless of individuals having off days (or weeks.) Even high performers can often afford to be reactive.
The stark reality of being an early-stage founder is that nothing happens if you don’t make it happen. Further, when you take your foot off the gas, you don’t stop. You roll backward. No experience pushes you harder to learn and grow fast. And there is no more gratifying work than taking the wild ideas in your head and, with complete freedom, turning them into real things that strangers find useful.
How do you see data teams (or the role of data scientist) evolve over the next 3-5 years?
An emerging “modern data stack” that combines data, analytics, and machine learning enables us to look back with zenith retrospective analyses. Companies can and should differentiate themselves by embedding that data into the projects they are working on.
While data science may have been considered a specialized role in the past, it is evolving to be the standard in building applications and driving business growth. Especially with greater accessibility we’ve gained with OpenAI. Because of that, data science is becoming even more democratized, with better tooling and techniques being developed daily.
What are some of your favorite data tools (other than Sumatra, of course)?
Like so many folks, I’m a big fan of dbt. Recently, duckdb. Rudderstack. Redpanda. AWS’ entire serverless stack: Dynamo, Kinesis, Athena, and now very excited about OpenSearch. If I have any hipster answer, maybe.
What led you to decide to build Sumatra?
I hate to see talented people wasting their time on useless stuff. I saw talent was no longer the bottleneck, but tooling was. While you always hear about big tech projects that build these things in-house, the impact is limited, and there is much wasted time and money. So many businesses have no interest in building large platform teams—nor should they. Nonetheless, have the need to.
The projects built for giant tech scale end up being awful from a UX perspective and serverless has made it possible for small teams to build it now. The agility and scale that Sumatra offers companies made it a no-brainer to build it.
Why was Lucas McGrew the right partner to build Sumatra with?
My advice to anyone looking for a co-founder is to, first and foremost, choose someone who is impeccably ethical. Nothing matters more than trust.
Lucas is the fastest learner of new technologies I’ve ever seen. He’s courageous. He has a can-do attitude. And he has a genuine desire for a more personalized future where online experiences feel intimate and businesses surprise and delight us with how well they know us.
Learn more about Lucas in his employee spotlight!
What is a core value of Sumatra that you are most passionate about?
Always be shipping. It describes our internal culture as well as our mission. It helps us stay nimble and embrace a mindset of continuous momentum and experimentation, which is incredibly important as a startup and in the use cases we are helping our customers tackle.
What part of your role do you enjoy the most and why?
First — working directly with customers. The possibilities that real-time data and machine learning create are endless, o we collaborate with our users to brainstorm and build new use cases. It also keeps challenging me to find interesting ways to incorporate our technology into every experience across our customers’ journey to help them understand the art of the possible in their applications.
Secondly — hacking the language. As a team, we’re constantly thinking of better and simpler ways to leverage the languages we use at Sumatra, SCOWL, and Python.
What are some of the current challenges/priorities you are focused on?
Our insight is recognizing that a wide variety of problems are, in fact, the same problem in a slightly different form. The main challenge we’re focused on right now is the communication challenge of helping people connect the dots between these seemingly different things. We’re doing that by putting out lots of examples and making them concrete.
You can see these in our “recipes,” where we break down these seemingly impossible things that are now possible with Sumatra into straightforward steps so teams can quickly implement and get value.
Learn more from Greg
We sat down with Sumatra’s Greg Kuhlmann to learn how he got his ML street cred from Apple, how he transitioned from data scientist to startup CEO and co-founder, and what happens when you take your foot off the gas (hint: you roll backward.)
How did you end up working on Fraud at Apple?
As I finished my Ph.D., a grad school friend reached out about an opening on his team at Apple, using machine learning to fight fraud on the App Store. I knew nothing about fraud — I had never owned an Apple product. But I guess I had enough ML cred to pass the interview.
Over the next eight years, I worked with an extraordinarily talented and humble group of people on projects with dizzying levels of impact. It’s harder to do ML at a company that genuinely cares about privacy. Fighting fraud at a company obsessed with user experience is even harder. For those reasons, I’m confident that Apple has the best anti-fraud data science team in the world, bar none. I feel grateful to have been a part of it.
How has the transition from a data scientist at Apple to Founder of a startup gone? Any lessons learned?
Large organizations, at their best, are these indomitable forces that move projects forward, regardless of individuals having off days (or weeks.) Even high performers can often afford to be reactive.
The stark reality of being an early-stage founder is that nothing happens if you don’t make it happen. Further, when you take your foot off the gas, you don’t stop. You roll backward. No experience pushes you harder to learn and grow fast. And there is no more gratifying work than taking the wild ideas in your head and, with complete freedom, turning them into real things that strangers find useful.
How do you see data teams (or the role of data scientist) evolve over the next 3-5 years?
An emerging “modern data stack” that combines data, analytics, and machine learning enables us to look back with zenith retrospective analyses. Companies can and should differentiate themselves by embedding that data into the projects they are working on.
While data science may have been considered a specialized role in the past, it is evolving to be the standard in building applications and driving business growth. Especially with greater accessibility we’ve gained with OpenAI. Because of that, data science is becoming even more democratized, with better tooling and techniques being developed daily.
What are some of your favorite data tools (other than Sumatra, of course)?
Like so many folks, I’m a big fan of dbt. Recently, duckdb. Rudderstack. Redpanda. AWS’ entire serverless stack: Dynamo, Kinesis, Athena, and now very excited about OpenSearch. If I have any hipster answer, maybe.
What led you to decide to build Sumatra?
I hate to see talented people wasting their time on useless stuff. I saw talent was no longer the bottleneck, but tooling was. While you always hear about big tech projects that build these things in-house, the impact is limited, and there is much wasted time and money. So many businesses have no interest in building large platform teams—nor should they. Nonetheless, have the need to.
The projects built for giant tech scale end up being awful from a UX perspective and serverless has made it possible for small teams to build it now. The agility and scale that Sumatra offers companies made it a no-brainer to build it.
Why was Lucas McGrew the right partner to build Sumatra with?
My advice to anyone looking for a co-founder is to, first and foremost, choose someone who is impeccably ethical. Nothing matters more than trust.
Lucas is the fastest learner of new technologies I’ve ever seen. He’s courageous. He has a can-do attitude. And he has a genuine desire for a more personalized future where online experiences feel intimate and businesses surprise and delight us with how well they know us.
Learn more about Lucas in his employee spotlight!
What is a core value of Sumatra that you are most passionate about?
Always be shipping. It describes our internal culture as well as our mission. It helps us stay nimble and embrace a mindset of continuous momentum and experimentation, which is incredibly important as a startup and in the use cases we are helping our customers tackle.
What part of your role do you enjoy the most and why?
First — working directly with customers. The possibilities that real-time data and machine learning create are endless, o we collaborate with our users to brainstorm and build new use cases. It also keeps challenging me to find interesting ways to incorporate our technology into every experience across our customers’ journey to help them understand the art of the possible in their applications.
Secondly — hacking the language. As a team, we’re constantly thinking of better and simpler ways to leverage the languages we use at Sumatra, SCOWL, and Python.
What are some of the current challenges/priorities you are focused on?
Our insight is recognizing that a wide variety of problems are, in fact, the same problem in a slightly different form. The main challenge we’re focused on right now is the communication challenge of helping people connect the dots between these seemingly different things. We’re doing that by putting out lots of examples and making them concrete.
You can see these in our “recipes,” where we break down these seemingly impossible things that are now possible with Sumatra into straightforward steps so teams can quickly implement and get value.
Learn more from Greg
We sat down with Sumatra’s Greg Kuhlmann to learn how he got his ML street cred from Apple, how he transitioned from data scientist to startup CEO and co-founder, and what happens when you take your foot off the gas (hint: you roll backward.)
How did you end up working on Fraud at Apple?
As I finished my Ph.D., a grad school friend reached out about an opening on his team at Apple, using machine learning to fight fraud on the App Store. I knew nothing about fraud — I had never owned an Apple product. But I guess I had enough ML cred to pass the interview.
Over the next eight years, I worked with an extraordinarily talented and humble group of people on projects with dizzying levels of impact. It’s harder to do ML at a company that genuinely cares about privacy. Fighting fraud at a company obsessed with user experience is even harder. For those reasons, I’m confident that Apple has the best anti-fraud data science team in the world, bar none. I feel grateful to have been a part of it.
How has the transition from a data scientist at Apple to Founder of a startup gone? Any lessons learned?
Large organizations, at their best, are these indomitable forces that move projects forward, regardless of individuals having off days (or weeks.) Even high performers can often afford to be reactive.
The stark reality of being an early-stage founder is that nothing happens if you don’t make it happen. Further, when you take your foot off the gas, you don’t stop. You roll backward. No experience pushes you harder to learn and grow fast. And there is no more gratifying work than taking the wild ideas in your head and, with complete freedom, turning them into real things that strangers find useful.
How do you see data teams (or the role of data scientist) evolve over the next 3-5 years?
An emerging “modern data stack” that combines data, analytics, and machine learning enables us to look back with zenith retrospective analyses. Companies can and should differentiate themselves by embedding that data into the projects they are working on.
While data science may have been considered a specialized role in the past, it is evolving to be the standard in building applications and driving business growth. Especially with greater accessibility we’ve gained with OpenAI. Because of that, data science is becoming even more democratized, with better tooling and techniques being developed daily.
What are some of your favorite data tools (other than Sumatra, of course)?
Like so many folks, I’m a big fan of dbt. Recently, duckdb. Rudderstack. Redpanda. AWS’ entire serverless stack: Dynamo, Kinesis, Athena, and now very excited about OpenSearch. If I have any hipster answer, maybe.
What led you to decide to build Sumatra?
I hate to see talented people wasting their time on useless stuff. I saw talent was no longer the bottleneck, but tooling was. While you always hear about big tech projects that build these things in-house, the impact is limited, and there is much wasted time and money. So many businesses have no interest in building large platform teams—nor should they. Nonetheless, have the need to.
The projects built for giant tech scale end up being awful from a UX perspective and serverless has made it possible for small teams to build it now. The agility and scale that Sumatra offers companies made it a no-brainer to build it.
Why was Lucas McGrew the right partner to build Sumatra with?
My advice to anyone looking for a co-founder is to, first and foremost, choose someone who is impeccably ethical. Nothing matters more than trust.
Lucas is the fastest learner of new technologies I’ve ever seen. He’s courageous. He has a can-do attitude. And he has a genuine desire for a more personalized future where online experiences feel intimate and businesses surprise and delight us with how well they know us.
Learn more about Lucas in his employee spotlight!
What is a core value of Sumatra that you are most passionate about?
Always be shipping. It describes our internal culture as well as our mission. It helps us stay nimble and embrace a mindset of continuous momentum and experimentation, which is incredibly important as a startup and in the use cases we are helping our customers tackle.
What part of your role do you enjoy the most and why?
First — working directly with customers. The possibilities that real-time data and machine learning create are endless, o we collaborate with our users to brainstorm and build new use cases. It also keeps challenging me to find interesting ways to incorporate our technology into every experience across our customers’ journey to help them understand the art of the possible in their applications.
Secondly — hacking the language. As a team, we’re constantly thinking of better and simpler ways to leverage the languages we use at Sumatra, SCOWL, and Python.
What are some of the current challenges/priorities you are focused on?
Our insight is recognizing that a wide variety of problems are, in fact, the same problem in a slightly different form. The main challenge we’re focused on right now is the communication challenge of helping people connect the dots between these seemingly different things. We’re doing that by putting out lots of examples and making them concrete.
You can see these in our “recipes,” where we break down these seemingly impossible things that are now possible with Sumatra into straightforward steps so teams can quickly implement and get value.