A Professor of Verification and Control in the Department of Computer Science at the University of Oxford, with 20 years in research. Earlier, he carried out research at Stanford University and at SRI International and was an Assistant Professor at TU Delft. He received a Laurea degree from the University of Padova and MS/Ph.D. at UC Berkeley.
How did your scientific involvement in AI & ML begin?
My academic background is in Applied Mathematics & Engineering. During a doctorate in the US, I became interested in Computer Sciences including areas such as AI and ML. I have been active in interdisciplinary areas across CS and Engineering ever since, with a focal interest in transferring my research into new technologies.
How did you get involved with Zinia?
I was approached by Aashutosh, CEO at Zinia. He saw an alignment between my ongoing interests in principled integration of model-based mathematical techniques with data-driven learning algorithms from ML, and Zinia’s vision. Our mutual interests coalesced into a natural collaboration and so, we began phases of critical research, development, and testing. We have been working together ever since.
What have you helped create at Zinia?
Over the past two years, I have helped the data science team in Zinia to develop the platform from concept to a scalable, adaptive product. We have formed a team of data scientists who are conducting product-focused research and technology transfer, so we can leverage state-of-art scientific results in AI and ML for business decisions.
What can Zinia do?
Zinia incarnates self-service AI: any user can seamlessly access Zinia as a tool to turn data into models and models into decisions and more generally into insight. Once you have uploaded your data via the Cloud, Zinia takes you on an automated, no-code journey from data explanation to predictive model selection, and finally to decision optimization. Also, expert data scientists can off-load automated tasks to Zinia, thus freeing their time for work that instead requires human expertise and ingenuity.
What do you think about the capabilities of AI-guided business decisions now and in years to come?
In the current world we can access ever more available data, so ML becomes a core component. At the same time, it is important to retrain principled approaches, particularly with see-through models built from data, as not all applications are about “big data”. Aspects beyond data crunching are also fundamental: guarantees for model transparency, certificates for decision optimization, and explainability for fairness and lack of bias.
