Ravi Patankar, Phd Email & Phone Number

Detroit Metropolitan Area

Click stream, Retail Predictive AI Models, GCP

Indianapolis, Indiana, United States

Develop IoT, data and analytics strategy, Communicate the strategy to executive leadership via presentations, white papers• Lead the team of 8 that developed transformer-based, deep learning predictive method for advance warning of compression cycle related field failures with 90% accuracy. See BLOG.• Invented energy saving HVAC control that factors in future weather in a closed loop MPC/RL. Energy savings reduce the carbon footprint of homes without sacrificing comfort, providing the thermostat an edge.• Developed crowd-sourced, accurate excess energy consumption alerts for individual homes as a precursor to system degradation and eventual faults. The alerts are based on physics inform statistical model.

Novi, Michigan

Setup Analytics group for model-based diagnostics product development at KPIT, Interface with potential automotive OEMs and Tier 1 customers to formulate the problems for customer pain points and solution proposals• Patented a graph-based approach to build a detailed diagnostic information model from engineering documents. Lead the team to develop and commercialize it into “Trace2Fix” product for maintenance and data collection.• Analyzed dealerships collected maintenance data using Trace2Fix to generate and benchmark performance metrics for the system in a live dashboard that provided the OEM previously impossible real-time intelligence.

Analytics COE design. Starting a dialogue with businesses and formulating end-to-end process optimization and analytics problems for the COE to target: retail, purchasing, inventory, oil and gas, refining, chemical process. Provide technical guidance to the team.• Established seabed subsidence predictive model using big data from sensor instrumented seabed. Proportion of seawater to gas increases significantly after seabed subsidence, actions could be taken in advance to avoid this expensive problem when subsidence prediction was made available to field maintenance.• Once a leak is suspected in undersea pipeline (KGD-6 oil and gas), it takes hundreds of miles of remotely operated underwater vehicle (ROV) navigation and the ROV carrying main ship to locate the leak. Using the data, we were able to locate the leak within 50 miles cutting the ROV search range in less than half.• Gradient search optimization of crude oil mix for maximizing refinery profit led to data-driven purchasing decisions. Chemical process optimization maximized yields.

Develop data mining and analytics functionality in Condition-Based Maintenance program• Lead data science activities to develop counter-intuitive field maintenance insights that solved urgent field maintenance issues for the GM IVHM program (patent applied)• Developed prognostic alert algorithms that were commercialized in Predictive Trend Monitoring and Diagnostics (PTMD) product• Developed aircraft fleet analytics web dashboard containing several ML algorithms collaborating with RapidMiner team and presented it organization-wide and used it for coaching and spreading data science awareness • OSST (Operations Support and Sustainment Technologies) 6.3 project– System architect and integration manager of 6 new predictive technologies on a helicopter HUMS box for a high-profile Army demonstration

Future Combat Systems (FCS)-Platform Soldier Mission Readiness System subcontract from Honeywell. Write research proposals to DoD, NASA, etc. Direct research for funded projects on Diagnostics, Prognostics and Health Management, Automated Maintenance Plan Generation and IETM assembly, and, Control systems.• Secured and executed several externally funded programs, patented inventions.

Advised 2 PhD and 2 MS students. Instructor: MEEM 4700 Dynamic Systems and Control MEEM/EE 5715 Linear Systems Theory and Design MEEM 6702 Nonlinear Systems Analysis and ControlResearch: Real-time nonlinear parameter estimation, Networked Control Systems, X-by-wire, Modelling fatigue crack initiation and small crack growth with state-space approach• Invented the controller for Electric Power Steering. Patent commercialized on several car models.• Several journal publications.

B Tech, M. Tech, Mechanical Engineering

Activities and Societies: Dynamic systems and control

Ph.D, Mechanical Engineering

Proposed a constitutive difference equation for the dynamic development of crack opening stress in fatigue crack growth. Developed.

Doctor Of Philosophy - Phd, Mechanical Engineering