At AscentCore, we’re always looking for ways to innovate and build our team’s capacity to serve our clients better. One way we’re doing this is with the creation of AscentCore Labs – a method for us to test and validate our ideas (and our clients’) to explore new opportunities through AI and other tools.
We recently sat down with AscentCore’s Chief Technology Officer, Cornel Stefanache, to learn more about the creation of AscentCore Labs and where it’s headed.
How did AscentCore Labs begin?
We first came up with the idea of Ascent Core Labs to have a ready team for helping our customers explore new business opportunities through AI, machine-learning, data visualization, and other tools. So the initial idea was to focus on proof of concepts that were developed really fast and to look more at the results, rather than the quality of the development. Essentially, we’re trying to validate either a client’s idea or our own ideas to make us even better at what we do and to be innovators when it comes to working with clients.
What are your current and longer-term projects for the Lab?
We’re currently working on research around explainable AI (AI predictions or insights that can be explained in a way that humans can understand) and we’ve had some interesting results. We haven’t been able to fully collect, analyze, and make sense of everything just yet, but we are pretty far in the process.
We’re also exploring the influence of synthetic data that can be used alongside training data to see if we can reduce the biased prediction of AI models. This is an automated process that is enabled through ML blocks, and will be released as an open source technology that facilitates our research. Ultimately, ML blocks will provide zero code infrastructure for organizations to adopt data science and machine-learning practices and integrate them into their ecosystems.
Looking ahead to the long-term, we’re following three vectors:
- Helping our customers explore new business opportunities
- Developing open source products that we use internally to validate ideas, but then will also be available to the community with an opportunity to adopt and improve our tools
- More research
What is the ultimate vision for AscentCore Labs?
Our primary goal is to gain a credible foothold in this space and end up with some great results from the tools we’re building. Ideally, we want to create an open marketplace to generate a standard in AI adoption with best practices in data analysis, data visualization, and prediction algorithms. We also hope to have a vibrant and active community contributing to the marketplace with components that can be easily integrated.
How do you measure success?
Success can be measured through adoption, ease of integration, and purpose. Building these tools will allow us to rapidly create an AI infrastructure for our clients and contribute our expertise and knowledge so clients can accelerate their growth, and hopefully their sales, too. It’s also an opportunity for our clients to quickly experiment with the adoption of these types of technologies.
How can interested individuals or communities join AscentCore Labs?
As I mentioned earlier, we want to create an active and vibrant community that is working to make improvements on what we’ve done so we’ll be publishing Docker images that folks can use to implement their customizations. We’re also going to create tutorials and a website that will house everything in one place. That way, anyone can come to the website, make requests for improvements, experiment, and contribute to problem solving.
We’re so excited about the possibilities of the AscentCore Labs and can’t wait to share what we’re learning. We think it will be a great contribution to the space of AI technology adoption and a way to solve challenges together. Keep watching this space as we release our research and for opportunities to contribute to the community!