Plus 9 years of experience in developing data analytics, data mining and modeling methodologies using JMP/R. Thought leader in applying advanced statistical and mathematical methods to engineering data analysis, sensor analytics and robust decision making Drove six sigma projects to improve product Quality/Yield which resulted in cost savings worth millions of dollars. Developed innovative features for trend diagnostics using Fourier based methods Developed statistical models to predict healthy die behavior which improved product quality and enabled wafer shipment. Expertise in multivariate analysis techniques (PCA, PLS) to analyze high dimension data for feature selection and yield prediction. Expertise in Clustering analysis and Classification Trees for diagnostics and root cause analysis
R, JMP, JSL, Excel, SQL, Spotfire
Statistical and Machine Learning methodologies as below
Data mining: Engineering Data Analysis, Unsupervised, and Supervised methods
Forecasting: Fourier analysis, MA, Exponential, ARIMA, etc.
Tree Methods: Classification and Regression Trees
Clustering Methods: K-means, Hierarchical
Multivariate methods: PCA, PLS
Univariate Analysis, Sensitivity Analysis
Certified Six sigma black belt
Member of Technical Staff September 2014 to CurrentYield Solutions Globalfoundries US INC
Requirement gathering from users Evaluation of requirements Research and develop algorithm and software solution to cater user needs Deployment and Optimization, maintenance of software solution Evaluate ROI for developed software solutions Key Projects (Software used R, JMP, Spotfire) Developed statistical solutions that help reduce product inline variability and prevent excursions worth millions of dollars.
Developed defect limited yield prediction using Poisson and Binomial model which helped engineers estimate yield outlook and enabled them to identify key defects, inspection layers which detract yield and fix those to improve yield.
Developed Kill Ratio methodology which helps engineers to focus on key defects of interest and fix them.
Developed parametric yield prediction model using techniques like PCA and PLS which help engineers to correctly target devices and identify key electrical parameters that modulate parametric yield.
Principal February 2011 to September 2014
Responsibilities Engineering data analysis, data mining for yield variation root cause analysis.
Yield variation analysis, Excursion prevention Six sigma consulting, Data analysis consulting Key Projects (Software used JMP) Developed and used advanced statistical methods (PCA, CART, ANOVA, Fourier Analysis) to solve yield variation issues costing millions of dollars.
Developed a signature based methodology which helped track and explain product yield variability.
Applied six sigma approach to solving yield excursions which resulted in saving to millions of dollars.
Defined DOE analysis methodology and coded it in JMP which enabled engineers to analyze their experiments and statistical interpret the results.
Developed outlier screening methodology based on wafer sort data to screen outliers as requested by customer which enabled shipment of wafers worth millions of dollars.
Data mining Engineer October 2008 to February 2011Micron Technology Inc
Use statistical methods for Yield improvement.
Monitor Yield trends, Electrical Test trends and support datamining for trend variations.
Key Projects (Software used JMP) Developed trend analysis software using Fourier Analysis method in R.
This analysis software was widely used by Yield and Integration engineers to monitor Yield, ET parameter trends and focus on key ET, Yield trending parameters.
Used K Means clustering in JMP identify spatial zones in wafer.
This methodology was used by device engineers to identify spatial regions on product to perform dose optimization.
Developed a conversion monitoring system which helped engineers to track process conversions responsible for Yield enhancement.
Quality Engineer December 2006 to October 2008Micron Technology Inc
Responsibilities Apply statistical methods to enable Fab engineers to solve their problems Key Projects (Software used JMP) Developed a Pace Diagnostics Tool which helped identify the bottle neck workstations for the Pace.
Used PCA in JMP to create a multidimensional metric which enabled disposition the bad quality wafers.
Developed a methodology using JMP to understand the impact of R2R on inline metrology.
The methodology resulted in offline analysis which helped process engineers to estimate the number runs to optimize process variable.
MS : Electrical and Computer Engg Dec, 06Portland State University － OR, USAElectrical and Computer Engg Dec