Open||Source||Data Reframing Machine Learning and AI-Assisted Development with Jorge Torres
This episode features an interview with Jorge Torres, Co-founder and CEO of MindsDB. MindsDB is a virtual AI database that works with existing data to help developers build AI-centered apps. In 2008, Jorge began his work on scaling solutions using machine learning as the first full-time engineer at Couchsurfing, growing the company from a few thousand users to a few million. He has also served a number of data-intensive start-ups and was a visiting scholar at UC Berkeley researching machine learning automation and explainability.
In this episode, Sam and Jorge discuss the inspiration and challenges behind MindsDB, classic data science AI versus applied AI, and time series transformers.
-------------------
“So much data in the world is time series data, so much data. Even data that people don't know is time series, it's time series. So long as it’s moving over time, it is time series data. Whether you store it or not, that's a different thing. For having a pre-trained model on time series data, it even enabled the fact that you don't have to store all the historical data. You can just take the model and start passing data as it comes through, and then you get out the forecast. So you don't even have to have the historical data. All you need to have is the data at that given instance, and you can pass it to the model and you get an output. It's mind blowing.” – Jorge Torres
-------------------
Episode Timestamps:
(05:20): The inspiration behind MindsDB
(10:20): Classic data science AI approach vs. applied AI
(22:09): What open source data means to Jorge
(28:51): What excites Jorge about Nixtla and time series transformers
(37:07): A question Jorge wishes to be asked
(40:20): Jorge’s advice for the audience
(41:38): Backstage takeaways with executive producer, Audra Montenegro
-------------------
Links: