Teikametrics

Exploring and developing a new SaaS platform fueled by data engineering

Icon of a triangle with an exclimation mark in the middle

Challenge

Explore complex product idea while hiring a team

Icon of a treasure map with four dots and an X

Solution

Design best practices, data engineering, Ruby on Rails, Scala

Icon of a location marker with five dashes coming out from the top

Outcome

Product-market fit, $9M ARR, upskilled engineering team

Video of Aatish talking about why he reached out to thoughtbot

Challenge

Explore an entirely new product direction that uses Machine Learning

Aatish Salvi joined Teikametrics as CTO to take on an ambitious plan he had formulated with CEO Alasdair McLean-Foreman: explore an entirely new product direction that uses Machine Learning to help small to medium-sized online sellers compete in the market. In addition to the technical challenges, Aatish needed to develop his engineering team and hire principal engineers at the same time.

In the video above, Aatish talks about why he reached out to thoughtbot to help him explore the concept with design thinking, architect their data pipeline, and ship an MVP. In the end, thoughtbot not only gave him the bandwidth to focus on hiring, but developed his existing team's capabilities and culture for future success.

Quote from the Teikametrics project

thoughtbot left me with a much more seasoned team than I had started with, a better architecture, and a much clearer roadmap. Even to this day, the DNA from thoughtbot’s early engagement lives through the company. They really set us up for the success we are today.

Aatish Salvi
CTO, Teikametrics

The tech behind the solutions

When Teikametrics reached out for help with their platform, thoughtbot initially built out a rapid MVP using Ruby on Rails.

As customers started getting accepted into the system, it became clear that more firepower was needed on the data side.

In order to keep up, thoughtbot built a separate service to process data from Amazon. For this service, Scala, Akka, RabbitMQ, and Postgres were used to build a lightweight but scalable data platform for Teikametrics. By utilizing distributed data streams, the platform can break down a company's entire advertising history in minutes, providing continuous recommendations to sellers. Because the stream processes data in constant memory and applies backpressure, massive influxes of data won't overwhelm the system, and additional data can be processed faster by adding new workers to the cluster.

Quote from the Teikametrics project

If you’re a starting CTO with a very large mandate and you are in this market where hiring is difficult and slow, it’s easy to end up compromising on hiring because you’re under so much pressure. Your better option is to go to thoughtbot and get those highly seasoned engineers in the building early. Get them to ramp up your existing team while you find the exact matches for who you want your principal engineering team to be.

Aatish Salvi
CTO, Teikametrics

What does success look like for your project?