In the News: Startup Daily covers Relevance AI’s $4M AUD on AusBiz

  • Interviewed by Startup Daily on Aus Biz
  • We discuss our vector based platform and our $4M AUD raise
  • Key Staff: Daniel Vassilev & Jacky Koh
min read

Have you seen our latest interview by Startup Daily on AusBiz? Relevance AI’s founders Daniel Vassilev and Jacky Koh were interviewed by Simon Thomsen to talk all about our search discovery & vector experimentation platform.


Simon Thomsen: Now let’s meet another startup and a couple of co-founders. Relevance has introduced a powerful developer-first, Vector platform to help developers do more with unstructured data. The company has also just closed a $4 million round to develop it. To tell us more, founders Jacky Koh and Daniel Vassilev, join us now. Jacky, Daniel, great to have you in the studio. Now you’re going to have to take me through a little bit of explaining what’s going on here because it sounds really cool, but I’m still trying to wrap my brain around it in terms of how you guys are tackling data, Jacky. 

Jacky Koh: Yeah. So as you mentioned, Relevance AI, it’s a experimentation-first, Vector platform, and we help data scientists, developers to pretty much utilize and rapidly experiment with vectors to create AI features with unstructured data. So what has happened here is that vectors and probably what a lot of people are questioning about what are vectors. 

Simon Thomsen: Bingo. 

Jacky Koh: Yeah, yeah. Vectors, essentially, is a way to represent unstructured data in a numerical format. It’s actually a format that a lot of the top tech companies, such as Google, Pinterest, Spotify, actually use in a lot of their popular products.

So Google search, for example, they actually vectorize every single text, in every single page, and also every single search query you make to actually be able to pretty much match the query against the specific results. And vectors are a great way to pretty much turn unstructured data into that format that computer can understand. But also even better, it’s actually able to solve similarity and relevance problems, hence the name, Relevance AI. 

Simon Thomsen: Ah now, I’ve got it. The light bulbs coming on for me. Daniel, you’ve got backing from Insight Partners, a US VC. They normally don’t go into early stage companies. So this is a fantastic validation of what you are doing here that they see that potential. 

Daniel Vassilev: Absolutely. So the partner that we work with, George Matthew, he used to be the CEO of Alteryx. And I think from his experience there working with data platforms and analytics or structure data, he saw how much of an opportunity there was in the unstructured space. And I think what we are delivering to the market, we’re very early in the market. And we’ve got, I think, a unique perspective on how we can provide people with the best sort of functionality when using vectors.

So it’s absolutely great validation. And I think it was a testament to them understanding how much this is needed, how much companies can benefit from actually using vector to process and manage unstructured data. And I think it just shows how much potential there is. It shows that every single business can benefit from better analyzing unstructured data.

80% of business data is unstructured. It’s things like images, PDF documents. And so we can all benefit from using this sort of technology. It’s just more about having the tools that enable us to use it. So we’re very glad to have partnered with Insight Partners, and we’re really fortunate of what we’re going to build together with them and the other investors that we have on the team. 

Simon Thomsen: And this seems to be one of those extraordinary moments where machine learning enables this possibility that wasn’t previously there. What led you guys down this path? You’ve got a great team. You all interested in that data science space and sort of suddenly banged heads together and thought, “Hey, here’s how we can tackle the problem of vectors.” 

Jacky Koh: Yeah. So the idea if I want to really think back, it’s probably through pretty much learning data science and really looking at the key parts of it. But what really came about of pretty much Relevance AI was actually…

Before Relevance AI, I was consulting for a lot of different large corporate companies, and they have a lot of unstructured data, especially in the format of text. And they weren’t really utilizing this. It was pretty much free form, text fields within a SQL table, or even a CSV. So what vectors actually enable is to make that all possible to and quickly analyze it.

So when we saw that opportunity, our team… I spoke to Dan who I’ve worked with pretty much since high school. We’ve built apps to together. And I was like, “Look, have a look at this. This is really cool. You can pretty much search using text to actually find images.” You’re not even tagging the images. You’re just vectorizing text and vectorizing the images, and it’s matching like that. So it’s purely machine learning based. So that was really cool.

When I showed him it, he was like, “Yeah, this is really cool.” So we started building the company, started building the team… When we pretty much founded a company as well, we tried to get it into the hands of customers as early as possible. And yeah. 

Simon Thomsen: That is absolutely amazing. So Dan, the $4 million, what does it mean for the business, the plans? 

Daniel Vassilev: Yeah, well the $4 million really has enabled us to grow and scale the business to the level that we really need to build this sort of functionality. So we managed as a fairly small team of four to five people over the last year to get to a stage where we had our product into customers’ hands and being used.

But we realized that if we wanted to deliver to the businesses that we deliver to today, we have customers ranging from scale-ups to ASX as companies. That we need to have a larger organization, both from being able to who constantly innovate and push the boundaries of what’s possible with vectors. But at the same time, deliver a product that has all the functionality that we want it to have, but also can deliver the sort of security and availability that our product can do.

We’ve tripled our head count since we’ve raised the round. So we’re currently around 18 and going to grow to the about 25. And with those people hopefully deliver our products to even more businesses throughout Australia and over the world actually. 

Simon Thomsen: Well guys, thanks for coming in. Thanks for teaching me a little bit about vectors. Now I can explain it to other people as well and absolutely amazing space. No doubt we’ll be talking to you again sometime soon. Really appreciate your time. 

Jacky Koh: Thank you. 

Daniel Vassilev: Thank you so much. 

In the News: Startup Daily covers Relevance AI’s $4M AUD on AusBiz
Benedek Zajkas
December 21, 2021
Interested in harnessing vectors to untap your data's potential?

Our vector experimentation platform will help you analyse your data immediately.

Talk to our team of data science experts