PROFESSIONAL SUMMARY:

• Data Scientist with 10 years of experience in Data Science, Mathematical Modeling, Simulation, design of experiments, Data Acquisition, Data mining, Statistical Analysis, hypothesis testing, linear and nonlinear predictive modeling, Machine Learning, Model validation, Optimization theory, State Estimation, Kalman Filtering, Observer Design, Feedback and Advanced Controller Design and Simulation. Applied knowledge in the fields of Automotive emissions, Biomedical, Chemical and subsea Oil & gas.
• Has PhD in chemical engineering - thesis involving Data driven modeling, Advanced controller design and Simulation.

COMPUTER SKILLS:

Simulation packages: MATLAB, SIMULINK, Aspen Plus, and AutoCAD

Data analytic tools: SQL, SAS, Tableau, R, Java, Python (Scipy, Pandas & sklearn)

Big data tools: Hadoop - Hive, Hbase, Pig & Spark

PROFESSIONAL EXPERIENCE:

Graduate Research Assistant

Confidential, TX

Responsibilities:

• Drug therapeutics & disease modeling and evaluation of interactions between the genes and proteins of gene regulatory network.
• Performed clinical data processing, cleaning, exploratory analysis, problem/question specification (training and test set), algorithm development, model selection, hypothesis testing, clustering, dimensionality reduction, development of input/output models using autoregressive (ARX) structures, parameter tuning, final model selection, testing (residuals vs fitted, residuals vs leverage, confusion matrix) and report writing using Confidential .
• Accessed large and unstructured datasets with hadoop streaming and pydoop, performed mapreduce jobs, stored and retrieved data from relational (MySQL) and nonrelational databases (Hbase), Developed models using Python modules SciPy, Scikit-learn and custom modules for ARX structures.
• Analyzed higher order nonlinear and hybrid Pharmacokinetic / Pharmacodynamic models and implemented describing function based approximating methods to study their interaction, and occurrence and stability of limit cycles.

Postdoctoral Researcher

Confidential, Houston, TX

Responsibilities:

• Developed data driven models, prediction methodologies for clinical decision support and designed advanced feedback controllers for automated critical care system.
• Problem definition/ solution formulation; developed protocols using design of experiments with Minitab for optimal clinical data collection; supervised data acquisition for noise and error free collection; data ingestion, exploratory analysis, feature creation, selection and evaluation using Hadoop Sqoop, Hive and Hbase
• Separated train and test data; Developed algorithmic supervised machine learning methods (parametric/non-parametric algorithms, regression, support vector machines and neural networks); built dynamic models using Python Pandas, SciPy and sklearn modules; performed regularization, cross-validation and sensitivity analysis.
• Designed, simulated and implemented a robust PID and model predictive controller (MPC) for the closed loop fluid and drug therapy. Presented and reported the data analysis and performance of the controllers to doctors using Tableau dashboards.
• Automotive emission control treatment project to develop lower order component models. Analyzed emissions data for hydrocarbon emissions and NOx gases; Developed reduced order combustion models for three way catalytic converter (TWC); Combined the TWC model with UEGO and HEGO sensor models and performed simulation using MATLAB and validated with real time sensor data to test the model fidelity; performed parameter sensitivity analysis.

Postdoctoral Researcher

Confidential

Responsibilities:

• Performed detailed analysis of various wireless networked control applications, system level performance i.e.
• Building Automation, Machine Control and Wireless tracking systems.
• Analyzed data from wireless sensor networks, classified the data to identify the lag, latency and missing data (due to packet delay & loss).
• Applied supervised (regression, classification, support vector machines) and unsupervised machine learning algorithms (clustering, dimensionality reduction, neural networks), residual analysis, variability investigation.
• Wireless networked control system co-design in the presence of different communication and routing protocols.
• Kalman filtering, Estimation for missing data. Event driven PID, Optimization based Event driven control, hybrid MPC Control.
• Prepared project proposals in collaboration with universities and industries to obtain funding from national and international funding agencies.

Research Consultant

Confidential

Responsibilities:

• Reviewed modeling and simulation of multiphase flow in pipelines. Developed mechanistic models using both theory and data from the literature.
• Solved the stiff system of DE's using numerical methods. Simulated the mathematical models using MATLAB and SIMULINK for flow regime Identification and predicted flow regime transitions with varying flow rates and PVT (Pressure, Volume and Temperature) behavior.
• Carried out simulations for flow regime predictions with varying rate of mass transfer between phases in horizontal, vertical pipelines and risers.
• Designed and solved the constrained optimization problem of flow rates in Christmas tree with reduced slugging while accommodating varying pressure drops in pipelines and well bores.
• Comprehensive review of Selective Catalytic Reduction (SCR) for mobile applications.
• Modeling, effect of temperature in various reaction pathways, ammonia slip, ammonia storage capacity of catalyst, catalyst aging and ammonia & NOx sensors.
• Developed control oriented models from the theory and by using data from literature to calculate reaction kinetics.
• Simulated the models and designed control framework for urea dosage and ammonia slip.

Graduate Research Assistant

Confidential

Responsibilities:

• Designed a laboratory experimental process station (using AutoCAD) with two highly nonlinear and multivariable processes, commissioned, and stabilized.
• Developed mathematical models, simulated, derived closed form solutions for nonlinear predictive models, and performed design and simulation of advanced controllers.
• Performed factorial and response surface based design of experiments using Minitab for acquiring optimal and informative data from processes for parameter estimation purposes; statistical analysis of data using SAS and SQL.
• Structure selection for nonlinear block oriented models;
• Hammerstein and Wiener models with linear dynamic and static nonlinear parts of different order;
• Developed closed form system identification techniques for nonlinear multivariable processes in continuous and discrete time.
• Validated the identified models in real time experiments.
• Simulation and tuning of nonlinear model predictive controllers based on block oriented Hammerstein and Wiener models.
• Designed and installed a laboratory process control setup.