With $8.8M Raised, Is This The Ultimate Machine Learning Tool?

Scott Clark

In Episode #705, Nathan interviews Scott Clark. He’s the co-founder and CEO of SigOpt, a Y-Combinator and Andreessen Horowitz backed, optimization as a service startup. Scott has been applying after-learning technologies in industry and academia for years. He holds a PhD for applied mathematics and an MS in computer science from Cornell University and a BS degree in mathematics, physics and computational physics from Oregon State University. He was chosen as one of Forbes’ 30 under 30 in 2016.

Famous Five:

  • Favorite Book? – The Hard Thing About Hard Things
  • What CEO do you follow? – Marc Andreessen and Ben Horowitz and Phil Knight
  • Favorite online tool? — Gmail and Slack
  • How many hours of sleep do you get?— 8
  • If you could let your 20-year old self, know one thing, what would it be? – Scott would tell himself that it doesn’t get easier, so set up habits and processes to make things sustainable when you have the time and ability to do it because that will definitely help once things ramp up

Time Stamped Show Notes:

  • 00:44 – Nathan introduces Scott to the show
  • 01:25 – SigOpt is optimization as a service
  • 01:27 – SigOpt helps companies build different, complex AI and machine learning pipelines
  • 01:41 – SigOpt is a SaaS model and the subscription is based on the number of models per month
  • 01:53 – Pricing starts at $2500 a month and enterprise starts at $10K a month
  • 02:13 – Average monthly RPU
  • 02:33 – SigOpt usually engages at the executive level
  • 02:38 – People wanted to use AI for their businesses but couldn’t find the right person to do the work so they go with SigOpt
  • 03:23 – One of SigOpt’s client is Prudential
    • 03:31 – Insurance companies are augmenting their traditional methods to the new data that is being collected
    • 03:48 – As their data increases, the need for the best possible performance increases
  • 04:26 – What SigOpt does is different from the traditional machine learning as a service company
  • 04:41 – Scott shares a specific example of how SigOpt works with credit card companies
    • 04:44 – Fraud detection has been around for decades
    • 05:28 – SigOpt fine tunes different knobs and levers in the configuration parameters that makes the machine model work
  • 06:15 – SigOpt focuses on black box optimization
  • 07:45 – SigOpt relies on the domain expertise of the person at the specific firm to build a deep learning model
  • 08:31 – SigOpt applies an ensemble of global optimization techniques to the problem so they can efficiently configure the system
  • 09:20 – SigOpt suggests different curvatures
  • 09:58 – SigOpt has raised $8.8M to date
  • 10:30 – SigOpt never sees the underlying data
  • 11:11 – The entire system is designed to be hands-off
  • 11:43 – SigOpt was launched end of 2013
  • 11:51 – Number of paying customers is around a dozen
  • 12:18 – Average MRR
  • 12:25 – SigOpt prefer annual deals
  • 12:54 – No churn yet
  • 13:14 – Team size is 13
  • 13:35 – The capital raised was spent on the team and the enterprise sales efforts
  • 13:54 – 3-4 of the team are in sales
  • 14:05 – CAC
  • 14:22 – They sometimes visit their customers
  • 14:55 – Investors like to make big bets on the new technologies
  • 15:33 – The goal for the series A money
  • 16:33 – Average expenses
  • 17:40 – The Famous Five

3 Key Points:

  1. The need for AI and machine-learning is growing fast and there’s not enough people who are qualified to develop these products.
  2. The headcount can eat up most of a company’s expenses—especially in the technology industry.
  3. Optimization services make a business more efficient leading to a less to none churn rate.

Resources Mentioned:

  • The Top Inbox – The site Nathan uses to schedule emails to be sent later, set reminders in inbox, track opens, and follow-up with email sequences
  • Klipfolio – Track your business performance across all departments for FREE
  • Hotjar – Nathan uses Hotjar to track what you’re doing on this site. He gets a video of each user visit like where they clicked and scrolled to make the site a better experience
  • Acuity Scheduling – Nathan uses Acuity to schedule his podcast interviews and appointments
  • Host Gator– The site Nathan uses to buy his domain names and hosting for the cheapest price possible
  • Audible– Nathan uses Audible when he’s driving from Austin to San Antonio (1.5-hour drive) to listen to audio books
  • Show Notes provided by Mallard Creatives