About BlueGen.ai
- Founders: Edwin Kooistra & Iman Alipour
- Founded in: 2022
- Employees: 8
- Money raised: €500,000
- Ultimate goal: To become the authority on synthetic data
AI systems are gaining momentum. With that comes an increasing demand for data. These intelligent models need to be trained, after all. However, data is often privacy-sensitive, which prevents companies from freely using or sharing datasets. The tech platform BlueGen.ai has an algorithm that fixes this problem. They use their software to create synthetic datasets that cannot be traced back to an individual. This eliminates the need to take privacy regulations into account. In this instalment of start-up-of-the-day, founders Edwin Kooistra and Iman Alipour talk to us about their AI system and synthetic datasets.
What is BlueGen.ai?
“Digital companies have to rely on data. It is their life blood. They need that data to be able to innovate. We see more and more datasets becoming available. Companies are also making more important decisions on the basis of data analytics. Except that they are struggling with privacy restrictions. We make that data more accessible so that they can manage their business better.”
Kooistra: “Our AI software learns from the original dataset and then generates a new version that resembles the original. It contains all the connections, but no privacy-sensitive or traceable data. At present, companies are still very restricted in what they can do with their datasets due to privacy reasons. While there are technologies out there that anonymize the data, they are not satisfactory. They destroy the correlations between data points. Our artificial dataset does allow companies to use that data to train their own models. Or to share them with other parties.”
How can you guarantee that data privacy is safeguarded?
Kooistra: “We have developed a mathematical model based on differential privacy that prevents data taken from a dataset from being traced back to individuals. That model has also been tested by the Centrum Wiskunde & Informatica (Center for Mathematics & Informatics, CWI). We also work with decentralized learning. Such datasets are not shared with our platform. The data remains with the companies. In fact, our technology is comprised of two components. Namely, a discriminator and a generator. The discriminator converts the data into usable statistics and sends it to the generator that then turns it into a synthetic dataset. That is, they do this without having seen the original data.”
How did you come up with the idea for BlueGen?
Alipour: “I got in touch with Delft Enterprises to see how I could be of assistance to start-ups coming out of TU Delft. As a matter of fact, the first conversation I had was with professor Lydia Chen. She told me that she was working with synthetic data. During that conversation with her, that idea already grabbed my attention right off the bat. The product was already fairly well developed by then, but they still needed people like us who had experience in setting up companies. We had those discussions back in July 2022, and BueGen.ai was born on 21 September. We are now several pilots further down the road and we have talked to dozens of companies that have expressed interest.”
How did the first couple of months go?
Alipour: “We’ve just started and we’re bursting with energy. With every day, we are achieving more success. Plus, we are gaining a lot of traction from the market that motivates us to keep on going.”
”Kooistra: “It is all progressing beyond expectations. By that, I am particularly referring to how people know how to find us. For example, we are being invited to give presentations and attend a start-up event like Slush. It was there that EDF approached us, one of the largest energy suppliers in Europe, because they were interested in our technology. This doesn’t normally happen within two months of starting up a company. It really is a bit of a roller coaster at the moment.”
Was it difficult to attract investment?
Alipour: “We haven’t sought investment yet. We are in a position to fund the pre-seed ourselves, with support from Delft Enterprises. However, we are currently in talks with a number of investors, although we don’t need funding at the moment. We are mainly interested in conversations and contacts to help us build our network and orient ourselves further. We are not ruling out seeking funding in a year from now, but that depends on how fast we grow.”
What makes your product better than what is already out there on the market?
Kooistra: “We’ve built up businesses before and served in several leadership positions. In this respect, we are different from other start-ups.”The technology was developed by TU Delft, where the latest research developments could be quickly applied to the product. This keeps us ahead of the market. This currently translates mainly into the accuracy and privacy of the synthetic data we generate, which is crucial for its success and adoption. In addition, decentralized learning allows us to roll out more securely and more quickly.”
Alipour: “Our technology stands out from others thanks to our mathematical model, which enables us to guarantee that privacy is safeguarded. As it happens, we have also learned that not all synthetic data is by definition secure. In some cases, it can still be traced back to the source data. This is something we are able to resolve with our model.”
What is the ultimate goal of BlueGen?
Kooistra: “We firmly believe that data innovation should be made possible in a way that is privacy-safe. The two do not need necessarily have to be at odds with each other. What is driving us right now is being able to really make an impact in socio-economic terms.”
Alipour: ” Our ultimate goal is to build a sustainable organization and become the authority when it comes to synthetic data. For one thing, I also think we need to provide some education. Companies need to learn that this technology is out there and how to harness it. So, I hope that we will eventually be seen as a trusted advisor.”
Related Posts
