Standard Data Founding Story


Every start-up has a story, and this one is just as interesting as the rest of them


Ever since I was 10 years old I’ve had entrepreneurial ambitions. This drove me to:

  • read extensively (no social life)

  • teach myself how to code (age 14)

  • release my own iPhone App (age 18) — had a few customers in Europe 🤠

  • major in Physics (get really good at solving problems) and, therefore, meet Thomas Duquemin.

  • Read WSJ every single morning in college.

  • invest >95% of my net worth into the stock market while in college (almost all TSLA — when everyone said it was going to $10 💀)

  • Apply machine learning methods to Physics research. A nuclear physicist PhD and I built an ML model together, this was my first foray into ML and he taught me everything I needed to enter industry.

  • Join a ML consultancy firm when I graduated in 2019 to end up leading multiple computer vision projects for the TSA and DHS.

2 Weeks Before COVID-19 Quarantine Begins…

My stocks were doing great and I was in a strong financial position to jump; Therefore, I quit my job and went straight to my attorney’s office in Austin to form a new entity — Peterson AI Inc. because I didn’t yet have a name, or even really know what I wanted to do: I just knew that I wanted to work on AI, build SaaS products, and help to empower the ML industry by making the tech more accessible.

Why Standard Data?

Since being introduced to machine learning by the postdoctoral nuclear physicist I shadowed at Texas A&M, I’ve been curious about it. Upon entering industry and leading a few computer vision projects, I quickly realized that the field is as fragmented as it is high-potential: There’s simply too much variation in approach, as opposed to software engineering where, for example, every repository more or less is styled the same: In other words, there are standards. As of Standard’s founding, there weren’t any concrete standards in the ML industry, and that issue still remains; However, I do believe OpenAI is taking steps in the right direction.

I’ve long been a big John D. Rockefeller (Sr.) fan — highly recommend reading the Chernow bio 💯. I liken the current situation with ML to the oil refining issues of the late 19th century. Before Rockefeller stepped in and “Standardized” the quality of kerosene, the industry was suffering immensely: Buildings were blowing up due to impurities in the kerosene — not good for an industry trying to takeoff. 🛫

Data is the new Oil.

Like oil, data is something that’s to be collected, refined, and turned into high-value outputs. Oil was the energy that powered the industrial age, and I hold that the same is true for data; Namely, that it’s going to be a larger driver of the information automation age that we are in now.

COVID-19 Is Upon us.

I had just received Standard’s EIN from the IRS when COVID lockdowns began and the world economy ground to a halt. There was no telling what would happen, but I remember thinking to myself: “Alright, well, no going back now” .

Of course I knew how to code, knew ML, but had never synthesized a full-fledged SaaS product — There is a lot more to this than meets the eye! I had no idea where to begin, but I did: I just started building SaaS products.

A Bunch of Duds.

I don’t like using the word failure because of the finality associated with it: If you stop trying then yeah, sure, it’s a failure. But if you never stop, then those “failures” are actually iterative steps toward success.

We started building and never stopped — no matter the impediments.

  1. Spikit — Machine Learning Dataset Service // Product failure ⚰️

  2. Zipit — AWS S3 Parallelized Zipfile Unzipper Service // Market failure ⚰️

  3. Motion — Notion page to website service // Product and Market failure ⚰️

  4. NotionMail — Notion page to email service // Product and Market Success ✅. Thomas joined at the start of this project, that’s probably why it’s done so good.

  5. Notion2Email — Integratable Notion page to email service // Product and Market Success ✅

  6. ThinkMyThoughts — Thought summarizer // soft-launched

  7. LitCommit — GPT3-based git commit message generator // soft-launched, great response so far.

Standard is a product company doing consulting, not a consulting company trying to build products.

There is a difference, believe me.

We gleaned important lessons from each one of these products. Not only has our team gotten absolutely deadly when it comes to SaaS development, but these product experiences have only bolstered our consulting side. No other consulting firm in existence began as a bootstrapped start-up, we did. I sunk everything into the company, and know the pains of limited runway, the importance of moving quickly and not wasting time on anything, especially development.

What are the future plans for Standard Data?

Chiefly, make AI more accessible by increasing the pace at which AI products can be developed and delivered to market. We will be delivering those products ourselves, and helping others do the same.

In addition, we are creating a community of AI people called Weekend AI Builders, with the intent of enabling people of all technical/data backgrounds to become world-class AI SaaS developers — in a weekend.

Dillon Peterson