TL;DR: I learned how to work at scale, had relative freedom, and got a decent amount of money as a bootstrapped solo entrepreneur. However, I found it difficult to do truly impactful things without a team/tribe.

Jul 19, 2019


Overview of the company

My company (Loki.ai) provides data APIs, dashboards, and iframes to media companies. It scrapes unstructured, publicly available data, and converts it into data feeds and articles in close to real time.

Some applications of these include air-quality monitoring, traffic monitoring, election dashboards, and monitoring of economic indicators.

An air-quality monitoring iframe that we provided to a media client in India

An air-quality monitoring iframe that we provided to a media client in India

Usage of the Loki dashboard for video creation and election analysis by a large media company in India

Usage of the Loki dashboard for video creation and election analysis by a large media company in India

At the same time, it also provides analytics APIs and data warehousing for a smaller subset of clients.

The road to setting up the company

I’m a 27 year old self-taught programmer. I launched a t-shirt business (with ~$20k in annual revenue) in high school, and then launched a failed analytics product for manufacturing companies while in college.

After finishing college, I launched another failed Edtech startup. After I ran out of money, I tried to get better at data science and freelanced for media companies.

Then, I worked as a Data Scientist for a (now-defunct) personalization startup for 8 months. You’ll see a pattern here. I hardly saw any tech success — though got progressively better with the experiences.

Running Loki.ai

I quit my job in December 2015, and started freelancing for media companies again to understand the industry better. Worked for cheap and tried to spend as much time in newsrooms as possible.

In the meanwhile, I also launched a news site called The Broadline in 2016 (which has now been shut down) to experiment with alternative ways of creating a CMSbetter analytics, and better recommendations.

An early iteration of the analytics dashboard developed for The Broadline

An early iteration of the analytics dashboard developed for The Broadline

Once it became clear that there was interest from multiple media companies in my services, and that I could automate some of the work that I was doing, I decided to incorporate a company to try and scale the process in September 2017.

The company focused only on the narrow problem of converting fast-changing, unstructured, publicly available data to automated visualizations and stories. I started with tracking economic indicators, air quality indicators, and automated traffic monitoring.

My very first product was a bunch of iframes for real-time pollution monitoring in India — launched in November 2017. Those did okay — getting 40k+ users in the first 10 days. This wasn’t particularly impressive, but was a useful proof-of-concept. The lessons learnt from the process also allowed me to improve other products and APIs.

First 11 days of the pollution iframes — our first product

First 11 days of the pollution iframes — our first product