Data Python

BHASHMI FATNANI

313-***-**** *******.*@************.*** linkedin.com/in/bhashmi-fatnani github.com/bhashmifatnani EDUCATION

Northeastern University, Boston, MA Dec 2020

Master of Science in Information Systems GPA: 3.6

Relevant Courses: Data Science, Deep Neural Networks & AI, Bigdata Intelligence and Analytics, Cloud Computing Symbiosis Institute of Technology, Pune, India Nov 2014 Bachelor of Technology in Computer Science Engineering Relevant Courses: Data Structures & Algorithms, Database Management System TECHNICAL SKILLS

Languages: Python, Java, C++, R, SQL, HTML/CSS

Libraries: Keras, Seaborn, Scikit-Learn, NLTK, Flask Databases: MySQL, Microsoft SQL Server, PostgreSQL Cloud: GCP, AWS, Docker, Jenkins

Machine Learning: Random Forests, PCA, SVM

Deep Learning: CNN, RNN, LSTM

PROFESSIONAL EXPERIENCE

Data Science Intern Intralinks, MA, USA Jan 2020 – Aug 2020

• Implemented NLP models such as LDA, LSA and LDA2VEC to perform topic modeling on dataset of documents

• Leveraged hierarchical clustering to optimize brute force duplicate finding by 75%

• Deployed the duplicate finding algorithm on AWS lambda

• Designed and developed an application using Flask to visualize word embeddings Software Engineer Syntel Pvt. Ltd., India Jul 2014 – May 2017

• Built modules using AngularJS by keeping MVC concept intact in Agile client project

• Performed exploratory analysis on health care data to extract features for predicting risk

• Managed team of 3 to develop business modules and ensured data availability to stakeholders

• Worked closely with product managers, designers and customers to deliver impactful product initiatives RESEARCH EXPERIENCE

Broad Institute Mouse Brain Mapping Python, CNN, TensorFlow, Google Cloud Dec 2019

• Improved test accuracy by 13% by experimenting with combinations of various data augmentation techniques

• Performed distributed data acquisition and auto stitched portions of high-resolution images to build the training dataset

• Used anti-aliasing to resize the high-resolution images to 512*512 PROJECTS

New York Taxi Fare Prediction Python, Pandas, Matplotlib, Google Maps API, forecastio Kaggle Competition

• Implemented advanced geographic mapping techniques and geocoding to build spatial visualizations in Jupyter Notebook

• Used Linear Regression, Random Forest Regressor and KNN to predict fare prices with baseline RMSE score of 4.73

• Improved RMSE to 4.3 with hyperparameter tuning using Randomized Search CV Image Classification & Hyperparameter Optimization Python, Numpy, CNN, Transfer Learning Personal Project

• Reduced training time by 25% using a bilinear algorithm to reduce image resolution of ImageNet to 64x64 pixels

• Increased dataset by factor of 6 by leveraging data augmentation to create synthetic images for target dataset

• Achieved 89.03% test & 93.1% training accuracy using an 8-layers CNN with transfer learning Uncovering sentiments using EDGAR datasets Python, NLP Academic Project

• Scraped and performed sentiment analysis on earning call transcripts from EDGAR for 10 different companies

• Processed data using NLP data pipeline, created BOW, GLOVE, and word embeddings models

• Used Amazon, Google & Watson API to obtain Sentiment score and Normalized score to create a baseline accuracy of 72%

Contact this candidate