| | | 20,000+ Fresh Resumes Monthly | |
|
|
| |
Click here or scroll down to respond to this candidate Candidate's Name
PHONE NUMBER AVAILABLE | EMAIL AVAILABLE | LINKEDIN LINK AVAILABLE
SUMMARY
Experienced Data Science professional with 2+ years of industry expertise in data analysis, visualization, ML and full-stack
development, seeking opportunities in Data Science and AI/ML fields.
EDUCATION
Masters in Artificial Intelligence Aug 2022 - Jan 2024
University at Buffalo, State University of New York, NY 3.3 GPA
coursework: Machine Learning (ML), Big data, Biometrics image analysis, Computer vision, Deep learning (DL), Pattern
recognition, Robotics algorithms, Numerical Math, Natural Language Processing (NLP), Data structures, Gen AI
PROFESSIONAL EXPERIENCE
Data Scientist - Jersey STEM, New Jersey Feb 2024 Present
Implemented LLama2 LLM model from Hugging Face using LangChain for text classification, significantly improving accuracy
through fine-tuning
Boosted AI model training accuracy by 12% through strategic data wrangling techniques and innovative pseudo -labeling
methods, validated by robust performance metrics
Managed and updated critical datasets to uphold data integrity and accuracy standards, essential for reliable model performance
Data Scientist - Favordrop, Buffalo, NY Jun 2023 Aug 2023
Developed sales forecasting models with improved prediction accuracy by integrating demographic and competitive datasets using
Python and SQL Server
Reduced manual work by 30% and saved operational time by implementing automated data retrieval and cleansing process
Developed an effective data pipeline for extract, transform, and load (ETL) processes
Performed A/B testing and used KNN, statistical analysis, and neural networks to predict client preferences with high accuracy,
enhancing customer satisfaction through personalized recommendations
Data Scientist - Wipro, India Dec 2020 July 2022
Developed and deployed ML models for sales forecasting, customer segmentation, and recommendation systems using
TensorFlow, PyTorch, and scikit-learn
Conducted extensive EDA using Python libraries (Pandas, NumPy, Matplotlib, Seaborn) and performed data cleaning,
normalization, and feature engineering to ensure high-quality input data
Created interactive dashboards and visual reports using Tableau to communicate data-driven insights and trends to stakeholders
Implemented model evaluation metrics (precision, recall, F1 score, ROC-AUC) and performed hyperparameter tuning with
GridSearchCV and RandomizedSearchCV for performance optimization
Graduate Teaching Assistant - University at Buffalo, Buffalo, NY Sep 2023 Dec 2023
Assisted a class of 40 students with ML concepts like Regressions, Clustering, NLP and Computer Vision
Developed auto-grading scripts for a faster grading and engaged in code review sessions
TECHNICAL SKILLS
Programming: Python (SciPy, Scikit-learn,TensorFlow, PyTorch), R, Java, JS, C, HTML, CSS
Databases: MySQL, SQLite, PostgreSQL, Relational databases(RDBMS), Azurely
Tools: VScode, Jupyter Notebook, Github, Linux, MATLAB, Tableau
Others: Docker, PySpark, Hadoop, Agile Methodology, MLOps
ACADEMIC PROJECTS
Customized Document-Based Chatbot
Developed a chatbot using Retrieval-Augmented Generation (RAG) with a pre-trained Large Language Model (LLM) to answer
queries based on specific documents
Utilized a vector database for efficient document indexing and retrieval
Implemented and fine-tuned the chatbot to provide precise and relevant responses, enhancing user experience with accurate
information retrieval
Water Management System
Developed a centralized database with a Flask API(REST) for sensor data communication and collection, and built a React
UI for real-time usage monitoring and billing analysis
Applied advanced time series analysis techniques (LSTM) on the collected data to predict future water demands, improving
resource management and operational efficiency
Secure Biometric Identification System
Annotated a custom full-hand image dataset, extracting finger bounding boxes
Trained a CNN model to detect and validate fingerprints by recognizing ridges and valleys
Achieved 75% accuracy and saved the trained model as a pickle file for secure biometric identification
|
