Raghavendra Guna Shekhar
San Jose, CA - ***** +1-669-***-**** ac8d33@r.postjobfree.com github.com/raghavgsk25 www.linkedin.com/in/raghavgsk25 Data enthusiast with 1+ years of experience in data analytics and machine learning (project and course work). Skilled in problem solving, programming machine learning, and statistics. Seeking a data scientist/Analyst role to upgrade my skills and knowledge in the field of Data Science.
EDUCATION
Master of Science in Electrical Engineering, San Jose State University. Aug ‘16 - May ‘18 Bachelor of Engineering in Electronics & Instrumentation, Anna University. Sep ’09 - May ‘13 TECHNICAL SKILLS
Programming & Scripting Lang: Python, R.
Databases: Oracle SQL, MySQL, NoSQL, HDFS, Hadoop, MapReduce, Spark SQL, Flume & Sqoop. ML: Neural Networks, Supervised Learning, Classification, Regression, Clustering, Natural Language Processing (NLP) Python ML Packages: Pandas, NumPy, SciPy, MatplotLib, Scikit-learn, OpenCV, Keras, Theano, Tensorflow-gpu Statistical Methods: A/B testing, Bayesian modeling, Data Mining, Dimension reduction, Regularization, Data visualization Other: Tableau, GIT, Flask, Google Analytics, Linux, HTML/CSS RELEVANT COURSEWORK
Probability & Statistics, Machine Learning, Neural Networks and Deep Learning, Statistical Learning, Computer Networks CERTIFICATIONS
• Machine Learning A-ZTM: Hands-On Python & R in Data Science: Udemy
• The Complete Oracle SQL Certification Course: Udemy
• Intermediate to Advanced Python for Data Science: Data Camp
• Data Cleaning and Visualization for Data Science: Data Camp PROFESSIONAL EXPERIENCE
Project Research Assistant San Jose State University, San Jose Jul 2018 – Present
• Building a tool for multiclass categorization of financial news with an improved accuracy of stock price prediction. Deep Learning Intern GYPSOID (Sports Analytics), San Jose Feb 2018 – Jun 2018
• Analyzed millions of data points from recorded sports videos to evaluate an athlete’s performance leveraging smartphone camera, hence reducing the equipment cost by 70%.
• Worked on different deep learning algorithms for object detection including Faster R-CNN, Mask R-CNN and SSD to optimize the average precision to 82%.
• Built deep learning-based recommendation system to enhance the coaching process from the available data points. Project Application Engineer Hitachi, India Sept 2013 – Jul 2015
• Involved in Engineering, Testing and Troubleshooting of PLC, DCS, SCADA & HMI system.
• Built an SQL Bridge for connecting SQL databases to Programmable Logic Controllers(PLC’s).[Ignition SQL Bridge Module]
• Improved the PLC’s average performance speed by 23% by freeing up the PLC memory and storing it in the SQL server.
• Conducted database work by writing queries to pull data from SQL server which serves as input for the PLC.
• Led the Customer Support Technical Department by giving training and 24/7 support to the On-site and Client Engineers, reflecting a 60% increase in the customer retention rate.
• Awards: SPOT RECOGNITION AWARD (Feb 2015)
ACADEMIC PROJECTS
Illumination Invariant Traffic Sign Detection using Deep Learning (Python,TensorFlow, Keras, OpenCv)
- Real-time traffic sign detection and recognition method that is robust against illumination changes by incorporating CLAHE
(Contrast Histogram Equalization) method.
- Comparative analysis of three different deep learning models implemented on Tensorflow and Keras framework reflecting Average Precision of >90 % on the LISA Dataset (US Benchmark Traffic Sign Dataset). Heart Disease Prediction using Linear Regression, Decision tree & Neural Net (Python, Scikit-learn, pandas, Tableau)
- Visualized the data and features involved to predict the heart disease & data cleansing using pre-processing techniques.
- Comparative analysis of 3 different algorithms to achieve an accuracy of >85% (Decision tree algorithm). Handwritten Number Image classification using Convolutional Neural Net (Python, Keras, PCA, LDA, CNN)
- Analyzed performance of neural network on the benchmark dataset of MNIST for recognizing hand-written digits.
- Performed dimensionality reduction on dataset using PCA, SVD and LDA techniques to enhance the performance by 15%. Iceberg or ship detection using neural networks (Kaggle Challenge) (CNN, OpenCV, TensorFlow)
- Built a Convolutional neural network algorithm that identifies if a remotely sensed target is a ship or iceberg from given satellite image database provided by C-Core.
- Different hyper-parameters were modified to achieve an accuracy of greater than 95 %. Bit Coin price prediction using LSTM (Python, Pandas, Keras)
- Developed a multivariate LSTM architecture using Keras to predict bit coin price based on a dataset from Kaggle.