SHIVAKUMAR VALPADY

adccy2@r.postjobfree.com linkedin.com/in/shivakumar-valpady-a24b34171 413-***-**** EDUCATION

MSc in Computer Engineering University of Massachusetts, Amherst GPA - 3.79/4.0 May 2020 B.E in Electronics and Communication Engineering Bangalore Institute of Technology, India GPA – 75.8/100 June 2018 Coursework: Deep Learning, Machine Learning, Computer Vision, Algorithms, Math Tools for Data Science, Internet of Things (Data processing from Sensors), DSP (Machine Learning Techniques), Probability and Random Process, Signals and Systems, Linear Algebra Certificate Courses: Neural Networks and Deep Learning, Improving Deep Neural Networks, SQL for Data Science, Machine Learning Foundations: Case Study, Programming for Everybody (Getting Started with Python), Python Data Structures, Introduction to Programming with MATLAB, Fundamentals of Digital Image and Video Processing. SKILLS

Programming Languages: Python, R (Data Visualization and Rshiny), MATLAB Packages and Libraries: PyTorch, Keras, TensorFlow, Scikit learn, OpenCV, Pandas, NumPy,dplyr, tidyr, lubridate, ggplot2, Rshiny. Databases: MongoDB, SQL

Operating Systems: Windows, Linux (Ubuntu), MacOS

PROFESSIONAL EXPERIENCE

INTERNSHIP:

Data Analyst Intern at Embue.co, Boston June 2019 – August 2019

• Querying data from MongoDB database, created a UI and Server for multiple charts using Rshiny package in RStudio. Data preprocessing and data cleaning was done using tidyverse and dplyr packages.

• The graphs and charts were displayed in the Rshiny app using ggplot2. GitLab repository was used to store, monitor and share the work of every member of the team.

• The raw data was transformed to meaningful data and deployed it to the customers as charts and tables using an AWS server. Project Intern at Preva Systems Pvt. Limited, Bangalore January 2018 – March 2018

• Designed and implemented Asset Management using Deep Learning Models such as LSTM and CNN. Integrated on Raspberry Pi using Python language to detect and recognize objects and alert asset theft in a facility.

• The live outputs were displayed on the monitor with probability value and a bounding box.

• The model for the same was presented to a panel consisting of Electronics Graduate students and Professors for evaluation. TEACHING ASSISTANT (University of Massachusetts, Amherst):

• General Physics Lab Spring 2020 – Present

Instructing for 4 sections as an Instructor. Other responsibilities included demonstrating procedures, leading discussions and grading student lab report and presentations.

• Digital Signal Processing (DSP) Fall 2019

Responsibilities include conducting presenting lectures, Doubt Clearing Sessions, grading assignments and grading exam papers.

• Electricity and Magnetism Lab Spring 2019

Instructed for 2 sections as Lab Assistant. Responsibilities included demonstrating procedures, leading discussions and grading student lab reports and presentations.

ACADEMIC PROJECTS

Lane Detection and Classification using Deep Convolutional Neural Networks Fall 2019

• Developed an end-to-end system based on Deep Convolutional Neural Networks for lane boundary recognition, clustering and classification.

• Lane Detection was also done using geometric approach like sliding window technique.

• The results of the two methods were compared for accuracies and performance. Image Captioning using Attention Models and BERT Embeddings Fall 2019

• Convolutional Neural Networks for image classification and object detection along with LSTMs for Language Modelling were used.

• To improve the accuracy, Attention models were used on the conventional model. The BERT embeddings are implemented for Encoder and Decoder of the transformers.

• The BLEU scores are calculated for the evaluation of the models. Recommendation Systems Spring 2019

• Content Based Filtering and Collaborative Based Filtering Recommender System was implemented to recommend Jobs in USA based on title, location and jobs of similar users using TF-IDF, cosine similarity and count vectorizer from scikit learn libraries.

• For a Song dataset, Top N-Trending Songs, Top N-similar songs and Top N-songs based on listen count was implemented using TF- IDF Vectorizer, Cosine Similarity and SVD.

Prediction of crop yield based on weather factors at a specific location Fall 2018

• Worked on implementing ARIMA model for prediction of optimal temperature for maximum yield of Wheat.

• Presentation of the same was carried out to show how the data was collected, processed and implemented. Predicting MBTA & PVTA Subway Arrival Fall 2018

• Worked on data preprocessing for information gathered from multiple sensors in subway which was later utilized to generate subway arrival estimation model.

Contact this candidate