My Data Science Portfolio

BERT Powered Resume Analysis App

Screenshot This project aims to develop a web application designed to assist individuals in enhancing their resumes. Leveraging advanced Natural Language Processing (NLP) techniques, the application analyzes resumes, suggests improvements, and provides answers to user queries based on the resume content. App Link - Hugging Face Link

Technologies Used: Python, NLTK, Spacy, StreamLit, Transformers (BERT).
Link to Project: GitHub

Michigan Lake Precipitation Prediction

Screenshot2 For my INFO6106 neural network course, my team and I built a ConvLSTM with encoder-decoder attention mechanism on satellite imagery and meteorological data to predict snowfall precipitation for the next 2days, achieved model performance F-1 score of 94%.

Technologies Used: TensorFlow, Keras, Deep Learning, Image Data Processing, RNN-LSTM, ConvLSTM, Attention Mechanism
Link to Project: GitHub

Bio Signal Analysis for Smokers

Screenshot2 This project employs Random Forest Classifier, SVM, Logistic Regression to predict smoking status with a 84% accuracy(RF). Through in-depth data exploration, crucial features like gender and hemoglobin levels were identified as key influencers in the relationship between health indicators and smoking behaviors. This endeavor contributes valuable insights to enhance personalized health interventions and predictive analytics in healthcare.

Technologies Used: Exploratory Data Analysis, Data Visualization, Feature Engineering, Machine Learning, Outlier Detection, Model Evaluation using F-1 Score metric
Link to Project: GitHub

Hindi Text Generation

Screenshot3 Led NLP data preprocessing, refining Hindi text for optimal model input. Integrated FastText embeddings for semantic nuances and contextual understanding. Developed a Recurrent Neural Network (RNN) for coherent Hindi text generation through trained preprocessing data. Implemented dynamic text generation for versatile and effective content creation based on user input or predefined contexts.

Technologies Used: Python, FastText, TensorFlow, NLTK, NLP, RNN, Data Pre-processing
Link to Project: GitHub

Summary of Skills

Programming Languages: Python, Java, C++, JavaScript
Data Science: EDA, ETL, Hypothesis Testing, Supervised, Unsupervised ML, Deep Learning
ML Frameworks: TensorFlow, Keras, NLTK, OpenCV, Pandas, NumPy, Matplotlib, Seaborn
Database Management: Postgresql, MySQL, Oracle SQL, MongoDB
Tools: Git, AWS, CI\CD, VS Code, Jupyter

Experience

Software Engineer

Newgen Software

Mumbai, India

June 2021 - May 2022

Built a custom utility for Yes Bank account opening software application, utilizing Java, J2EE, reduced manual work by 20%
Developed a frontend portal to track status of all the pending files at the user department through reports, utilizing MS SQL database for backend
Added UI to existing module, designed a wizard for a form flow for user friendly experience, decreasing common user defaults by 25%
Optimized complex SQL queries, extracted discrepancies in records, and improved query execution time by 15%, optimized overall system response time
Implemented CRs (change requests) in process specific .war files, incorporated JavaScript for dynamic validations, resulting in enhanced form functionality
Collaborated with cross-functional teams, documented work, maintained code stability, corrected improperly coded loop, incorporated proactive problem solving of complex software issues efficiently

Education

Master of Science in Information Systems Northeastern University, Boston, MA Expected May 2024

Bachelor of Engineering in Computer Engineering Mumbai University, Mumbai, India June 2021

Contact

LinkedIn: Shruti Tambe LinkedIn
Email: Tambe.sh@northeastern.edu
GitHub: GitHub Profile

Acknowledgments

Inspiration - Chip Huyen Chip Huyen

This portfolio is licensed under the MIT License.