Michael Rinehart Email & Phone Number

San Jose, California, Us

Securiti is the pioneer of the Data Command Center, a centralized platform that enables the safe use of data and GenAI. It provides unified data intelligence, controls and orchestration across hybrid multicloud environments. Large global enterprises rely on Securiti's Data Command Center for data security, privacy, governance, and compliance.Securiti has been recognized with numerous industry and analyst awards, including "Most Innovative Startup" by RSA, "Top 25 Machine Learning Startups" by Forbes, "Most Innovative AI Companies" by CB Insights, "Cool Vendor in Data Security" by Gartner, and "Privacy Management Wave Leader" by Forrester.For more information, please visit securiti.ai and follow us on LinkedIn.

Mountain View, California, Us

Led a team of data scientists and software engineers to develop the key data-science products for Symantec CloudSOC, the leading security stack for enterprise cloud applications.Led the design and development of CloudSOC Detect, an advanced anomaly threat-detection engine. It includes detectors that account for time/day, access patterns, data volumes, data content, and other dimensions of behavior. The system is capable of continuous learning, high-fidelity modeling based in behavioral research, and a custom form of multi-level learning that operates on arbitrary distributions at very high speeds for day-0 detections and peer-group comparison.Led the design and development of the algorithms for CloudSOC’s cloud-based DLP system ContentIQ. ContentIQ leverages computational linguistics and a novel contextual-analysis engine to provide best-in-class detection of compliance documents (PII/PCI/HIPAA/…) as well as high-precision detection of challenging document types such as design documents.Designed the machine-learning pipeline for ContentIQ’s machine-learning based document fingerprinting. Several approaches in computational linguistics were needed to allow the user to provide only a handful positive training documents and yet receive high-precision detections.Led the initial design and development of the automation technologies behind CloudSOC StreamIQ. StreamIQ translates network traffic into logs of user actions. The data science team developed a set of human-machine interactive tools backed by machine-learning algorithms. The pipeline significantly increases the rate of CloudSOC cloud-app coverage. It further leverages unsupervised learning techniques to determine when retraining is required.Initiated and guided an effort to develop machine-learned classifiers to aid in the research of cloud application capabilities. Resulted in a significant increase in the breadth and accuracy of coverage.

Basking Ridge, Nj, Us

Developed a generalized linear model solver with L1-L2 regularization over MapReduce using the Alternating Direction Method of Multipliers over L-BFGSExtended Google’s PLDA C++ codebase to a smoothed variant of Link LDA for topic modelling under multiple topic sources (start-to-finish POC in several days).Developed large-scale scripts to process and filter HTTP traffic (Scala + Scalding) and location data (Python + Pig).

London, Gb

Algorithms lead for Phase II of DARPA’s BLADE program to develop a real-time jamming system to counter new and dynamic wireless communication threats. Determined initial module-level algorithms, system-wide algorithm to coordinate module interaction, and functional architecture. Managed a research team to develop algorithms and worked with a software team to develop system. Lead presentation efforts to DARPA.Learning lead for Phase II DARPA/NGA URGENT program to detect man-made objects in 3D-LiDAR data. Managed a research team to develop algorithms to determine optimal object detection parameters. Developed automatic and user-interactive approaches. Latter approach allowed user to find near-optimal detections in seconds.

Co-founded company to develop a bio-impedance system to measure muscle quality. Co-wrote SBIRs and wrote our first winning NSF Phase I SBIR proposal.Developed a new technique for fitting non-convex Cole-Cole models to bioimpedance data by splitting into two convex optimizations. Fitted model discriminates good/poor measurements. Interpolated complex impedance from fit shown to benefit data analysis.Invented a new tetrapolar electrode configuration for measuring anisotropic tissue impedance. Leverages symmetry to reduce electromagnetic measurement parasitics at high frequencies.Co-designed early experiments and performed early analyses relating measurements to muscle quality and demonstrating the benefits of EIM over certain functional tests for determining the progression of neuromuscular diseases.

Cambridge, Ma, Us

Researched the volatility of a DC power grid model. Applied spatial-Fourier analysis techniques to bound volatility as a function of a stochastic demand model.Supervised a masters-level graduate student researching the control of hybrid electric vehicles using map data. Guided research directions and implementation considerations.Researched the stability and efficiency of real-time electricity markets under different pricing, prediction, and consumer/producer interaction models.Provided the basis topic for masters student's thesis regarding dynamics in social networks.

Bank of America (Boston, MA), Oct. 2010-Nov. 2010. Examined the complexity of computing statistics related to loan defaults and developed algorithms to estimate the default distribution.Infolenz, Inc. (Cambridge, MA), April 2010-June 2010. Developed classifiers for predicting physician malpractice risk.

Ph.D., Electrical Engineering

M.S., Electrical Engineering

B.S., Electrical And Computer Engineering (Concentration In Computer Science)