Sanmitra Bhattacharya is excited to use data and AI to create new products and services that have never been possible before.  

Can you share the most interesting part of your career journey?

As an artificial intelligence and machine learning practitioner, I have been fortunate to work on problems that have direct impact on human health. Earlier in my career I worked on public health problems such as identifying factors that influence engagement of federal health agencies with the general public using machine learning models, natural language processing (NLP)-based methods for surveillance of health beliefs in the population, etc.

AI is transforming medical research, human health, and our ways of managing health care, and I am excited to contribute to this journey of “AI for good.”

In my work at life sciences organizations, I have worked on problems ranging from applications of AI in protein engineering for gene editing-related cancer research and treatment, to tracking nascent adverse events that are reported in online health forums and social media.

In my recent work at Deloitte I am leading data science teams to build models to identify fraud, waste, and abuse from health care claims. AI is transforming medical research, human health, and our ways of managing health care, and I am excited to contribute to this journey of “AI for good.”

What excites you most about working with data and AI?

Despite the deluge of data and the widespread adoption and applications of AI, data and AI tools are still in their infancy. I’m excited to be part of the journey of this rapidly evolving data and AI tool landscape and the potential to solve problems that have been intractable for decades. I’m also excited about the potential to use data and AI to create new products and services that have never been possible before.

Describe an interesting project that you have worked on.

Fraud, waste, and abuse (FWA) in federal health care programs cost taxpayers tens of billions of dollars while putting beneficiaries’ health and well-being at risk. One of the most interesting projects that I have been working on recently is the application of knowledge graphs and graph neural networks (GNNs) to identify FWA in health care claims.

Our team is following the latest research on this area, and developing cutting-edge AI methods are unique and novel in this domain.

While traditionally this type of transactional data is treated as flat data structures without interconnections within and across them, more recent progress in research on knowledge graphs and GNNs has led us to finding these interconnections and identifying FWA that could not be done with traditional rules and simpler analytics. Our team is following the latest research on this area, and developing cutting-edge AI methods are unique and novel in this domain.