The Art of Making the Possible Practical: Introducing Immuta Research

Drs. Joseph Regensburger,  Alfred Rossi, and Stephen Bailey trickle in on a Monday morning. In between sips of coffee, the conversation flows from familiar domains of life — traffic, kids, the events of the weekend — to heady technical topics. Frenetic whiteboard scribbling punctuates the discussion which, today, includes how to achieve a mathematical guarantee in a certain privacy model, and how to quantify model failure probabilities under perturbation of the training set. Meet our research team. Together, they ensure that Immuta is on the cutting edge of data engineering, analytics, and information privacy. Their work enables our customers’ data science teams to work with all of their organization’s data without being hamstrung by complex regulations or information governance policies.

The team’s current efforts center on developing the Immuta platform’s ability to quantify and manage algorithmic risk, as well as facilitate privacy-preserving machine learning workflows. To do this, they partner closely with Immuta’s Customer Success and Legal Engineering teams to understand the current state of technical, regulatory, and ethical challenges, and develop use cases within the analytics community. They identify specific areas in which Immuta helps solve these challenges for organizations and develop state-of-the-art solutions which can be implemented into the Immuta product. Far from being an academic silo, the research team partners with every internal division of Immuta, and externally with our partners and customers to make sure the platform is innovative, practical, and useful.

A strength of the team is that our researchers have diverse experiences and backgrounds:

  • Dr. Joseph Regensburger leads Immuta’s Research group: Prior to joining the company, he was the Chief Scientist at Illumination Works, LLC and Principal Research Scientist at the Battelle Memorial Institute. He has directed research efforts characterizing airport security screening devices, engineering image analysis software, and developing Machine Learning algorithms for biological detection. He was recognized with Battelle’s Technical Achievement Award and Illumination Works’ Innovation Award. Joe received his Ph.D. in Physics from The Ohio State University, with research focused in Experiment High Energy Physics, specifically the detection of rare decays of D0 Mesons.

“I joined Immuta because we are addressing some of the underserved challenges in data science.  Ethics and privacy are vital concerns within data science and analytics.  Algorithms are being tasked with performing a greater number of tasks in the world.  We need to make sure those algorithms aren’t compromising privacy or reinforcing biases.  At Immuta, we are helping to solve these problems.”  – Dr. Joseph Regensburger

 

  • Dr. Alfred Rossi is a theoretical computer scientist and research scientist at Immuta, where he is focused on the company’s data privacy efforts.  Before joining Immuta, Alfred was a Senior Lecturer at the Ohio State University, consultant, and a Researcher at the Battelle Memorial Institute. As a consultant, Alfred helped guide a top 250 US website toward microservice architecture. He developed a reverse proxy for real-time traffic analysis, cryptographic interfaces to file storage APIs, and microservices for digital signing. At Battelle, Alfred received five outstanding performance awards over a four year period. His efforts spanned numerous domains, including: mathematical modeling, design, and management of HPC clusters, and evaluation of data interchange standards for security devices. Alfred holds a Ph.D. in computer science, and an MS in physics, from the Ohio State University.

“While I’m deeply excited by the promise of machine learning, I’m increasingly worried about a future where we yield evermore decision-making power to black boxes. How do we know if they are functioning properly? If they are fair? How can we be sure that their behavior does not reveal personally sensitive information about training data subjects? Today, while it seems like the rest of the world is off thinking about the possibilities of these black boxes, we are busy figuring out how to design effective guardrails. The decision to join was easy — who could say no to incredibly meaningful, exciting work with a wonderful group of talented people? I’m grateful to be a part of it.” – Dr. Alfred Rossi

 

  • Dr. Stephen Bailey is an applied data scientist at Immuta, where he analyzes product performance to drive the development of new features and improve existing ones. He also researches the analytic tools and use cases of prospective Immuta customers to build reference implementations demonstrating analytic and machine learning techniques. Prior to joining Immuta, he researched the brain basis of reading disability using magnetic resonance imaging. He is a former National Institutes of Health Research Fellow, an instructor for DataCamp.com, and has published papers in several academic journals. He holds a doctorate in neuroscience from Vanderbilt University and undergraduate degrees in chemistry and philosophy from Samford University.

“As someone who’s worked with sensitive medical data, I’ve seen how challenging it can be for enterprises to make resources both accessible and useful — and how the issues only compound when ML gets involved. At Immuta, I met an extraordinary team that is prepared to close the gap between performing responsible analytics in theory and in practice.” – Dr. Stephen Bailey

 

Keep an eye out for a post from our research team in early February about why aggregate-only queries provide a false sense of privacy, and how differential privacy provides more guaranteed privacy protections.

In the meantime, click here for a 1:1 demo to see how Immuta’s platform can help you rapidly personalize data access and dramatically improve the creation, deployment, and auditability of machine learning and AI.