SUMMARY:

• Data Scientist/Machine Learning Engineer with 10+ years of industry experience
• Develop and deploy machine learning models based on Artificial Neural Networks, Principal Components Analysis, Random Forest, Gradient Boosti ng, Partial Least Squares, k - means Clustering. Applications include Time Series Forecasting, anomaly detection, computer vision, sequence modeling, topic classification, sentiment analysis, Recommendation Engines, real-time process control/optimization.

TECHNICAL SKILLS:

Programming Languages/Packages: Python, R, Matlab, Keras, TensorFlow, Scikit, Numpy, Pandas, Matplotlib and Seaborn, Natural Language Tool Kit, Apache Spark, Hive, NoSQL

WORK EXPERIENCE:

Data Scientist/Machine Learning Engineer

Confidential, Santa Clara, CA

Responsibilities:

• Object identification and logo recognition in household images (AWS Marketplace model packages: GluonCV YOLOv3 and Sensifai)
• Automated parsing of PDF forms to extract key-value pairs/Named Entity Recognition (AWS Textract, Spacy)
• Call volume forecasting for customer service center (LSTM Sequence to sequence); Monitor news stories and associated sentiment to predict impact of external events (data breaches, natural calamities, etc.) on call volume to optimize call routing, resource scheduling and prepare response strategy
• Topic clustering/classification and sentiment analysis to summarize and obtain actionable insights into customer complaints/reviews (LDA2Vec, Natural Language Toolkit, Spacy)
• Real-time anomalous topic detection based on call/chat transcripts (LSTM)
• Recommendation engine for prediction of customer problems and possible solutions based on customer’s geo-demographic data and interaction history (Random Forest, Gradient Boosting)

Applications Engineer

Confidential, Methuen, MA

Responsibilities:

• Convolutional Neural Network model to predict the physiochemical properties of gasoline from hyper-spectral images
• Deep Reinforcement Learning to determine optimal control trajectory in closed loop sequential control for cost minimization
• Design of Experiments based on Latin Hypercube sampling to generate data and train a simulated model of the environment for Reinforcement Learning model
• Hyper-parameter search using Bayesian optimization
• Significant contribution towards the growth/success of a successful startup company (revenue realization to increase in revenue up before acquisition by Confidential )

Applications Scientist

Confidential, Woburn, MA

Responsibilities:

• Principal Components Regression and Partial Least Squares Regression to predict physical properties of oil and natural gas samples in real-time (PCR, PLSR)
• Confidential model for high-altitude aerial monitoring to detect and quantify natural gas pipeline leaks using images captured by a LIDAR/Tunable Diode Laser
• Weibull distribution fitting and Cox regression modeling of data collected from a ccelerated life tests, warranty/failure data (Reliability analysis, S urvival/Hazard R ate F unction)
• Interpret statistical methodologies and results to non-statisticians in a multidisciplinary research team/environment

Research Assistant

Confidential, Raleigh, NC

Responsibilities:

• Random Forest based model to classify gasoline samples into different octane ratings with high accuracy for compliance and reporting at retail sellers/distributors
• C lustering based sampling technique to address between-class and within-class imbalance to improve classification performance of Random Forest models in a highly imbalanced multi-class classification problem
• Cleaned and analyzed large and complex data sets
• Review, implement and modify statistical algorithms described in research papers or available under open source

Associate Software Engineer

Confidential

Responsibilities:

• Write and execute test case scenarios/scripts to validate business requirements for Siebel based customer relationship management application
• W rite test scripts for quality assurance of web based financial applications