Kaustubh Waghavkar

**** ***** **., ***# ***, Santa Clara- 95054 Home: 951-***-**** Email id - adci05@r.postjobfree.com SUMMARY:Emerging Software Engineer and Data Scientist with hands-on experience & innovative approach to building leading-edge analytical solutions. Leverage combination of Programming, Big Data, & Machine Learning (ML) models to deliver value-adding insights on real world data. EDUCATION

University of California, Riverside - Master of Science in Computer Science Expected: June 2020 Savitribai Phule Pune University (Formerly University of Pune) - Bachelor of Engineering: Computer Engineering May 2016 TECHNICAL PROFECIENCIES

Programming Languages: Python 3, C, C++, HTML, CSS, JavaScript, Object Oriented Programming & Data Structures and Algorithms Frameworks: PyTorch, TensorFlow, OpenCV, Pandas, NumPy, NLTK, sklearn, matplotlib, GitHub, Git, HDFS, Hadoop & Spark. Databases: SQL, MySQL 5.1, Cassandra DB & MongoDB Cloud: AWS (basics), Microsoft Azure & Google Colab Tools: PyCharm, Jupyter Notebook, Eclipse, SVN, WEKA, KNIME, Tableau, Apache Tomcat, MS Excel & MATLAB (R_2019) PROFESSIONAL WORK EXPERIENCE

TEACHING ASSISTANT University of California Riverside, CA 06/2019 – 07/2019 Delivered instructional support in high-volume academic environment. Cultivated professional peer relationships and mentored junior colleagues.

(CS 100) Software Construction - Taught implementation of various design patterns in C++ using Google Test Framework to reduce test time by almost 86% compared to traditional testing methodologies.

(CS 242) Information Retrieval and Web Search - Designed coding exercise to compare BM25 and TF-IDF algorithms. Results indicated BM25 algorithm works 60% efficient compared to TF-IDF in scoring and ranking documents.

(CS 161) Design and Architecture of Computer Systems - Conducted interactive sessions to enable deep understanding of system design and familiarity with processor components such as pipeline, caches, and registers. SOFTWARE ENGINEER, KPIT Technologies Pune, India 06/2016 – 07/2018 Automated data conversions with purpose-built Python Parser and Python Crawler resulting in streamlined workflows, increased database space and 20% profit for team. Process was put into production for enterprise-wide level use through agile software development. Designed driver score module prompting harsh acceleration and braking score on vehicle. This significant security feature update increased customer satisfaction and 30% direct profit gain for team. Received Employee of the Month award for achievement. Executed vehicle tracking program for fetching 75K Point of Interest (POI) via Google API for 8K latitudes and longitudes. Significantly reduced tracking redundancy by 25% and generated faster real-time updates. Created analysis and handling instrument for Production Databases supplying competitive edge to increase sales by almost 80%. Wrote debugger script to conduct in-depth analysis, troubleshooting, bug fixing, automation testing, deployment of production issues. Decreased redundancy by 85% with enhanced error handling mechanism and code-by-code optimization. SIGNATURE PROJECTS

Master’s Mini Project: 02/2020 – Present

Principal architect of Python Crawler to gather monthly data from rent.com, rental.com, and indeed.com. and store in MySQL server of home analytics group at University of California, Riverside. Deployed Support Vector Machine algorithm to formulate and train ML model in transforming data as input to analyze trends in sought-after property rental fees, specific job availability, and salary information to forecast future amounts. Master’s Thesis: 08/2019 – 02/2020

Directed algorithm implementation using Computer Vision and Machine Learning to identify forging behavior of bees for Entomology Dept. Research outcomes have achieved 68.8% accuracy rate. Project honored with Shipley-Skinner award.

• Algorithm used: Kalman Filter, Mean Shift Filter, Linear Support Vector Machine (SVM). Ecosia - The Climate Search Engine 01/2019 – 03/2019 Engineered platform with web interface to search inverted index of 5GB of twitter data via Twitter Streaming API. Displayed top outcomes utilizing algorithms that designated sentiment score for returning positive, negative or neutral results.

• Framework: Hadoop Database: MongoDB Algorithms used: TF-IDF, BM-25 algorithm. Application of Big Data Frameworks to visualize Martian chemical composition 01/2019 – 03/2019 Employed Big Data models in creating tool to evaluate spectrometer data extracted from NASA MAVEN and ISRO for predicting life on Mars. Designed Graphic User Interface (GUI) to plot results and estimate chemical composition of planet.

- Models: HDFS, Map reduce Data pre-processing: Pandas (Python) GUI: MATLAB Prediction of Mid-Term Elections 2018 using Sentimental Analysis of Twitter Data 09/2018 – 12/2018 Executed Application Program Interface (API) for processing Twitter Search API textual data to assign sentiment score to each tweet that helped in predicting the winning chances of a party before the mid-term elections of 2018.

• Algorithm accuracy resulted in 87.39% for Democrat and 79.20 % for Republican data respectively.

• Data pre-processing: Text Blob (Python) Algorithms: Naïve Bayes & Logistic Regression Algorithm Machine Learning boot camp 01/2017 – 04/2017

Designed customized Convolutional Neural Network (CNN) with four layers using CIPHAR-10 dataset using PyTorch.

• Successfully achieved 74.22% algorithm accuracy rate on test dataset on a GPU instance in AWS. IEEE CONFERENCE PUBLICATION 03/2015 – 05/2016

Developed a system to predict the depression level using humans body signals and recommend a suitable therapy. Link: Healing Hands for Depressed People (D-HH) through analysis of human body signals to predict the level of depression and recommendation of suitable remedy.