SOPHIA SUSAN RAJU

*** **** **** **, #***, Santa Clara, CA, US-95054 https://github.com/sophiasr92

+1-669-***-**** adhxs9@r.postjobfree.com www.linkedin.com/in/sophia-susan-raju Exploring Machine learning/Deep Learning opportunities with 4+ years of software programming experience. Skilled in machine learning, data analysis, problem solving and programming. SUMMARY

● Strong knowledge and skills in ML/DL techniques–Developed and implemented various ML algorithms using Python

● Developed models using Neural Networks for applications in Natural Language Processing (NLP), Computer Vision, Object Detection, Image Classification, Speech Recognition

● Experience and Excellent programming skills in Python and C/C++

● Expertise in ML libraries such as Scikit-Learn, Pandas, Numpy, Matplotlib, Seaborn and Deep learning libraries such as TensorFlow, Keras, Pytorch, OpenCV and Experience in Data Mining and Data Analytics.

● Expertise in Designing and Optimizing Multilayer perceptron, CNNs, Recurrent Neural Networks and LSTMs

● Implemented model optimization methods, quantization, pruning and clustering.

● Proficient in Algorithms, Data Structures and Object-Oriented Programing

● Strong knowledge in Advanced Computer Architecture

● Strong knowledge in RDBMS and SQL database

PUBLICATIONS

• Sophia Susan Raju, “Application of Noise to Avoid Overfitting in TCAD Augmented Machine Learning,” IEEE 2020 International Conference on Simulation of Semiconductor Processes and Devices.

• Kashyap Mehta, Sophia Susan Raju, et al "Improvement of TCAD Augmented Machine Learning using Autoencoder for Semiconductor Variation Identification and Inverse Design," IEEE Access. EXPERIENCE

Machine Learning Intern Intel Corporation May 2020 – Present o Analyzing Intel’s Xeon server CPU’s high-volume manufacturing test data and automating data analysis and visualizations using Python scripting.

o Developing ML algorithms to predict the test results as pass/failure and troubleshooting the same. ML Research Assistant San Jose State University Nov 2019 – Present o Applying ML algorithms for semiconductor defect identification and inverse design o Developing ML algorithms and models to predict various parameters of a semiconductor such as temperature, width of channel and work function using python for the experimental data provided. o Implementing Autoencoder, PCA analysis on the experimental data to remove noise for better prediction of values. Senior Project Engineer Wipro (Client - Cisco, Santa Clara) Sep 2014 – Dec 2018 o Worked on Einstein AI, the cloud platform for implementing Deep Learning for customer preferences. Einstein Analytics, a platform to perform data analytics and visualization o Led developing team of 4 members and built applications on Salesforce CRM platform using custom code and integrated with external systems using REST and SOAP APIs. Followed best practices, CI/CD using tools like Bamboo and Jenkins and experience with common bug, version control and build systems like JIRA and GIT. o Experience working in Agile and Scrum methodology Project Intern Indian Space Research Organization (ISRO) Jan 2014 – Mar 2014 o Designed a system to control the speed and position of servo actuator in the launch vehicle, which is based on a BLDC motor and the whole system was implemented in FPGA EDUCATION

M.S., Electrical Engineering (3.75/4.00) San Jose State University, USA Dec 2020 B. Tech, Electronics and Communication Engineering CUSAT, Kochi, India May 2014 PROJECTS

Model Optimization Using TensorFlow for Automatic Modulation Classification o Implemented post training quantization, quantization aware training, pruning, and clustering to compress the model. o Compared the performance of each of the methods. Applied the optimization for Automatic Modulation Classification problem.

Computer Vision: Autonomous Driving

o Predicting vehicle angle in different settings for the autonomous driving system. Participated in Kaggle competition o Developed an algorithm to estimate the absolute pose of vehicles from a single image in a real-world traffic environment using CenterNet as a baseline model which models an object as a single point which is the center point of a bounding box.

o Implemented various regularization methods such as data augmentation, early stopping, image preprocessing technique, GCN and tuned the various hyperparameters to improve the performance and reduce overfitting. Research on Quantized Neural Networks for Mobile Devices o Implemented lower bit quantization using TensorFlow lite on existing models o Quantized the neural network model for MNIST classification using post training quantization COVID19 Global Forecasting Forecast daily COVID-19 spread in regions around world Apr 2020 – May 2020 o Participated in Kaggle's Covid19 forecasting series in week 3, week 4 and week5 o The goal is to predict the confirmed cases and fatalities between April 15 and May 14 by region and identify factors that appear to impact the transmission rate of COVID-19. o Developed various models for prediction which include Polynomial regression, Random Forest, and XGB regressor Stress detection and Classification using WESAD Data set o Developed different classification models to classify the stress levels using the WESAD dataset. o Seven different models were developed and compared the performance. The developed models include Logistic Regression, LDA, KNN, Random Forest, Decision Trees, Support Vector Machines. o Initial models were fine-tuned to improve performance further. Used different feature extraction methods and obtained optimized features. Outliers were also handled which improved the performance of each model. Deep Learning Classification of Fashion MNIST dataset using DNN Image classification o Developed various deep neural network models by varying different hyperparameters to classify the 10 classes of images in FASHION MNIST dataset and compared the performance of each. o Implemented the project in Python using Scikit-Learn and TensorFlow o Each of the models was developed by varying the hyperparameters and observed the performance of each model Deep Learning Kannada MNIST using CNN in Keras Image recognition and classification o Developed neural network model for the classification of Kannada digits using CNN in Keras using Python. o Implemented Convolutional Neural Networks and classified the images of Kannada digits. Used various regularization techniques such as data augmentation, Dropout, Early Stopping, etc. and obtained an accuracy of 98.6 in classifying the images in the dataset.

NLP TensorFlow 2.0 Question Answering

o Developed neural network model to predict long and short answer responses to real questions in Wikipedia articles o Utilized neural networks to implement text summarization and information extraction. TECHNICAL SKILLS

• Tools: MATLAB, JupyterNotebook, PyCharm Eclipse, Jira, Git, Synopsys VCS, XilinxISE, ModelSim, JMP

• Programming Languages: Python, MATLAB/Octave, System Verilog, C/C++, MySQL, JavaScript, CUDA

• Operating Systems: Unix, Linux, Windows

RELATED COURSEWORKS

• Academic Courses: Deep Learning, Machine Learning, SoC Design, Advanced Computer Architecture

• Online Courses: 1. Machine Learning (Stanford University, Coursera), 2. Algorithms (Stanford University, Coursera) 3. Deep Learning (Coursera) 4. Python for Machine Learning and Data Science (Udemy) ACHIEVEMENTS

• Intel Recognition Award: For outstanding contribution in automating the data analysis of server data for the design and validation teams.

• Best Newcomer Award when joined the new team in Cisco

• Best Employee Award

• Rashtrapati Award (President’s Award for the meritorious service in Bharat Scouts and Guides Association)

Contact this candidate