OdinDB
A platform for data sharing that allows contributors to control their data and companies to collect it ethically.
The OdinDB wordmark, inspired by the Norse god of wisdom and thunder
Contributions
- Presentation Design (Figma)
- Public Speaking
- Project Management
Problem
Connor, Kunal, I realized the idea for OdinDB at the intersection of two predictions:
The proliferation of AI and ML technology will create a market for large, high-quality datasets to train these models. They will become increasingly expensive to source and time-consuming to vet for accuracy and formatting.
Just like people started caring about organic food in the last decade, the average consumer will become sensitive to where data AI/ML models are sourced.
Through its marketplace, OdinDB would have the potential to be the first mover in the market for ethically-sourced, quality data.
Explaining our go-to-market plan at the startup pitch competition where we had the idea for Odin.
Solution
How OdinDB works
Companies would post Kickstarter-like “data campaigns” where contributors (users) choose to submit their data.
Once the dataset reaches a desired size, OdinDB would clean and format the data (using machine learning) to incentivize companies to collect and purchase data through us. Our target market would be businesses without large-scale R&D departments, and they would be purchasing access to an instance of a dataset at that specific point in time.
To incentivize personal data contributions, 20% of the revenue generated from the sale of a dataset would be returned among all users who contributed to the dataset, and at any point users can query and remove their data from any dataset.
Feedback
Here’s what judges from Loftium, The Institute of Translational Health Sciences, and Pioneer Square Labs had to say:
Interesting idea. Like the concept of users opting-in to monetize their personal data. Pairing with data-intensive corporations/end-products super compelling for a scale perspective.
Vision of the product and why a user would sign up for Odin could have been communicated more clearly. Did not feel as though the topic related enough to the overall theme (ouch!).
I think the idea of fingerprinting all your data so you can see where it went is very interesting. What sort of data trail tracking is feasible? Could you use the aggregated trails to optimize data traffic resources and infrastructure?
Reflections
This was the first product or idea that I’ve been a part of creating that felt like it was commercially viable, which was really exciting.
Pitching is difficult and I will always speak slower than I think when I’m on stage. I’ve gotten marginally closer to understanding how difficult pitching an idea for funding is.
I believe we could have better allocated our time working on the actual pitch– of the 24 hours we were given, we used 10 hours to ideate and design and only 2 hours to rehearse our presentation.
Extras
The deck I designed: Link
OdinDB in “real life” 