YASH DESHPANDE

**** ********* ** ****, ****** Meadows Riverside, CA, 92507 +1-951-***-****

adbaxp@r.postjobfree.com

SUMMARY

● Proficient at Python programming, particularly in the Data Science area

● Logical thinker with a high aptitude for learning new technologies quickly

● Implemented several projects in the analytics space with quantifiable outputs that resulted in providing actionable insights

● Strong background in Computer Science with specific emphasis on Databases and Machine Learning

TECHNICAL SKILL SETS

Programming: Python, R, C, C++, Objective C, RESTful APIs Java, Swift5 Web Development: HTML, CSS, D3, Javascript, SQL, NoSQL: Mongodb Change control/Repositories: Git

Big Data: Hadoop, Spark, Pig

Data Science: Matplotlib, Seaborn, Tableau, Folium, Pandas, Keras, SciPy, numpy, Scikit-learn, TensorFlow, Pytorch

Visualization tools: Tableau, PowerBI

Github: https://github.com/YashDeshpande1771

CERTIFICATIONS AND ACHIEVEMENTS

● Certified course in Machine Learning through Coursera (96.5% GRADE)

● The publication “An Approach Towards Enhancement of Review Summarization Using Sequence Model And Word Embedding” won the best paper award at ICCUBEA-2019.

● Published research paper “Product Review Summarization With Feature Extraction and Opinion Mining” in IJSERT vol. 8 issue March 2019.

EXPERIENCE

Client: Linkedin (through Raja Software Labs) June 2019- July 2019 Software Engineer Intern

Raja Software Labs (RSL) is a software development consultancy, focused on mobile technologies.

● Extensive involvement in the complete iOS Development lifecycle.

● Responsible for performance optimization through internal restructuring of code. Migrated libraries of LinkedIn iOS application from Swift5 to Objective C to improve application loading time.

PROJECTS

High Relevance product feature extraction from Amazon.com’s product reviews This application allows the user to quickly summarize the key features of a product sold on Amazon by searching through the product review section.

Created unsupervised and supervised models to detect features in product reviews. 90% accuracy in the unsupervised method and 98% accuracy for the supervised method on the validation set.

● Used TF-IDF and Recurrent Neural Network to extract product features from Review text to improve Amazon’s Keyword extraction

● Used Word embeddings (GloVe) to extract topmost relevant features

● Used Aspect Based Sentiment Analysis dataset to generate an LSTM model to predict the sentiment of different aspects of products

● To perform extractive summarization used Google’s Sentence Encoder to generate a similarity matrix; using the TextRank algorithm, generated review summary.

● Designed a web application for a user to fetch and display a summary Synthetic Image Generation on MNIST Dataset using Generative Adversarial Neural Networks (GANs) and data augmentation.

Here the application will allow the user to detect Synthetic/edited images. The model was trained on a set of true images and the program is split into an image generator and a discriminator, Achieved 98.5% validation accuracy on the discriminant model for synthetic images generated by the generator model.

Adult Income Classification

To determine a person’s income based on census data

● Used PCA and correlation graph for data reduction to improve classification performance to 84%

● Did a comparative study of Machine Learning models (Artificial Neural Networks, SVM, Naïve Bayes, Logistic Regression, Decision Trees) to find the best fit. Random Forest seemed to be the most effective at this task.

Freshii: A Meal Box Ordering Web Application

Developed an end-to-end food ordering web application that allows users to select items from the menu and place an order.

● Conducted implementation in accordance with the eight golden rules of Human-Computer Interaction (HCI)

● Designed and developed UI and interaction modules in Bootstrap and Javascript Toxic Comment Classification

Implemented deep learning models to identify toxic comments on social media platforms

● Top 1% solution on Kaggle by combining LSTM and CNN with Max-pooling layer

● Preprocessed 150k rows using PySpark Framework

Stock Market Analysis on Time-series data

• Demonstrated correlation between CCI and stock prices.

• Developed an auto regression model (ARIMA) to predict stock prices. Operating Systems Capstone

Assignment to modify an operating system as per specifications

● Replaced XV6 OS Round Robin Scheduler with Stride and Lottery Scheduler

● Created a kernel thread in XV6 OS for a user program EDUCATION

Master’s in Computer Science, University of California, Riverside (Sept 2019-March 2021) GPA: 3.66. Relevant Coursework (so far): Advanced Operating Systems, Big Data Management, Data Mining Techniques

Bachelors’ in Computer Engineering, University of Pune (July 2015- July 2019) GPA: 3.5. Relevant Coursework: Database Management Systems, Data Analytics, Data Mining, Machine Learning, Software Engineering, Web Development, Engineering Mathematics (Linear algebra, Calculus, Probability, Statistics)

Contact this candidate