Description
Alan Shimel, Cofounder and CEO of Techstrong Group, interviews Selector’s VP of Product Management, Kevin Kamel about Selector AI’s Spring 2024 Release.
Video length: 21:20
Speakers:
Alan Shimel, Co-Founder and CEO of Techstrong Group
Kevin Kamel, VP of Product Management at Selector
Transcript
Alan: This is Techstrong TV. Hey everyone, welcome back to Techstrong TV. I’m excited to introduce a new company to our audience today. Especially given the current macro conditions in tech, it’s great to hear about new companies growing and solving both old and new problems. With that, let me introduce you to Kevin Kamel. Kevin is the VP of Product Management at a company called Selector. Kevin, welcome to Techstrong TV—it’s great to have you here.
Kevin: Oh, thanks for having me.
Alan: Pleasure. So Kevin, before we dive into Selector—and we’ll spend most of our time on that—I always like to give people a sense of who they’re talking to. You’re the VP of Product Management, so obviously you handle product-related aspects, but you weren’t always in that role at Selector. Tell us a little bit about your journey.
Kevin: Sure, it’s a bit unconventional. I actually started as a statistician at NASA.
Alan: Okay.
Kevin: Yeah, I was involved in a field called orbital debris analysis, and that’s where I got my start. From there, I transitioned into an incubator around 2001. Washington, DC, where I was based, doesn’t have a big startup scene, so the best way to get exposure was through an incubator hosted by the University of Maryland. Eventually, I found myself working at an email marketing company, which was part of my journey to Selector. Email marketing is a very statistics-based business, so my background was a good fit. We handled sending newsletters, figuring out who opened and read them, and we did this at a large scale. People were very interested in the numbers—who clicked, what they clicked on, and why. This led us to develop monitoring capabilities, where we had to track how long it took to send these newsletters to millions of people. People wanted them delivered at exactly 9:00 AM, so we had to monitor both the system infrastructure and the network to ensure we had the capacity to send all this email over the internet, which had modest infrastructure in the early 2000s. That’s how I got started. I had the good fortune of being at a company called MailerMailer, which exited in 2017. Afterward, I worked for a monitoring and observability-focused company, where I ran product for about five years before being introduced to the folks at Selector. And here I am.
Alan: Very cool. They weren’t calling it observability in 2017, though, right? That term must have come later during those five years.
Kevin: They were, though I’m not exactly sure when the term was coined—maybe around 2013 or 2014. But yeah, monitoring is sort of legacy now, and observability is the new term.
Alan: Yep, for sure. By the time people see this, I’ll probably be in Paris for KubeCon, which has become a big observability event with its focus on cloud-native. But let’s talk about Selector. You were recruited into Selector. What’s the story behind it?
Kevin: Absolutely. So, at a high level, Selector offers observability and AIOps solutions—a hybrid approach. Selector is a platform that transforms your operational data into actionable insights. We collect telemetry from your environment and apply AI and ML to capture the leading signals within your network, infrastructure, and applications. This helps identify emerging problems or issues that have already occurred.
Alan: Okay, if you don’t mind, let’s peel the onion back a little. What are you using to collect this data? Is it based on open telemetry, Prometheus, or something like Grafana?
Kevin: That’s a great question. At Selector, we have a phrase: “any data from anywhere.” And we mean it. Our platform uses a no-code/low-code approach to data collection. For the more technical folks, it’s a declarative ETL. You can define the inputs and outputs for a given telemetry interface without writing any code, which is important because there are so many protocols and technologies out there. It’s not uncommon to encounter proprietary endpoints without known interfaces or libraries, so our platform is designed to quickly integrate with any data source. This is especially important in large enterprises where proprietary tools may still be in use after a decade.
Alan: Excellent. The observability space has really expanded over the last few years, with AI and ML becoming almost mandatory to handle large data sets. But we’re also seeing a challenge with the sheer volume of data—it’s not just about analyzing it anymore. The cost of storage and computational power to maintain these data sets is almost prohibitive. We’re in an era where we can measure everything, but do we really want to?
Kevin: That’s an interesting point. As a former practitioner, I remember when a small group of people could understand what was happening with a few systems at any given time, and as long as you didn’t make any late-in-the-day changes, you were good. But that world is long gone. Storage costs are a major problem, and that’s why many enterprises are reassessing whether their observability platforms should serve as long-term historians or be more operationally focused. Selector is more operationally focused. We collect telemetry, analyze it, and send Smart alerts to your team that identify the root cause of a problem and its impact, allowing you to address it immediately. Unlike other platforms that may focus on warehousing data and providing dashboards for historical analysis, we prioritize actionable guidance for operations teams.
Alan: So, it’s not really a forensics tool but rather something that provides real-time advice.
Kevin: Exactly. Our audience is primarily operations teams, particularly network operations teams. While we can certainly support other use cases, we’re starting with network-driven engagements because that’s where our strengths lie. From there, we plan to expand into application operations and other areas.
Alan: Got it. And how is Selector offered? Is it a SaaS tool, traditional software, or something else?
Kevin: It’s flexible. Selector is a fully Kubernetes-based stack, which allows for various deployment options. Many of our current customers prefer on-premises deployment, but we’re moving towards availability in the Google Cloud Marketplace later this year. This will allow users to spin up an instance of Selector with just a few clicks.
Alan: That’s great. And how is it priced?
Kevin: I’ll leave that to the sales team!
Alan: Fair enough. Who is the target audience? Is it large enterprises, mid-sized companies?
Kevin: Our sweet spot is large enterprises to mid-sized companies. While we can handle small business use cases, that’s not our primary focus.
Alan: I know you recently upgraded some features. Can you tell us about that?
Kevin: Absolutely. We’ve integrated generative AI into our stack, which is a novel feature. It allows customers to use plain language to ask questions of the system and their telemetry, eliminating the need to learn a domain-specific language. We also introduced a feature called Retrieval-Augmented Generation (RAG), which leverages various pieces of information to provide comprehensive responses. For example, if a router fails, the system can look up relevant incidents and knowledge base articles, and even suggest potential fixes.
Another new capability is our root cause analysis, which uses causal ML to identify the precipitating event of an incident and order the subsequent events. This supports our smart alerting system, which provides operators with actionable alerts.
Finally, we have the ability to model an entire infrastructure, creating a digital twin of the network. This is especially valuable for large service providers, allowing for both real-time monitoring and capacity planning.
Alan: That’s great. Kevin, I just realized we didn’t mention the website.
Kevin: It’s www.selector.ai.
Alan: That’s S-E-L-E-C-T-O-R.A-I. Kevin, thanks for coming on and telling us about Selector. I hope this won’t be your last time here. Are you guys going to be at KubeCon?
Kevin: Not this year, but maybe in Salt Lake City in the fall.
Alan: Great, we might see you there. Kevin Kamel, VP of Product Management at Selector. Check out Selector.ai for AIOps and observability.