Zuber Resume Aug23
Zuber Resume Aug23
Skills Summary
• Python: NumPy, Pandas, Matplotlib, Plotly, PySpark, Sklearn, NLTK, spaCy, Flask, Pytorch, Optuna,SHAP.
• Databases: SQL Server, MongoDB , MySQL , Postgresql
• Azure Services: Azure Data Factory, Azure Databricks, Azure blob Storage.
• BI Tool: Qliksense , Power BI.
• Machine Learning and Data Science Skills: Regression, Classification, Clustering, GBDT, XGBoost , Random
Forest ,CNN ,Word Embedding, Statistics ,Feature Engineering , Feature Importance ,Hyperparameter Optimisation.
• Other Tech Stacks: Docker, Git , Github Action,Power Automate
Experience
Data Scientist May 2022 - Current
•
Piramal Pharmaceutical Solutions Mumbai, India
Customer Feedback Analytic’s:
• Extracted data from survey sparrow for each customer to employ BERT encoding for prediction of sentiments
increasing efficiency the overall process by 50 to 70%.
• Developed data pipeline to process, tokenize and loaded pre trained BERT model to obtain good precision,
recall and F1 score hence utilized 3 to 4 evaluation metrics
• Constructed NLP system with Pytorch, PySpark framework which automatically classified any number of
feedbacks within 120-180 seconds.
• Facilitated data driven strategies for enhancing customer satisfaction and issue resolution based on sentiment
analysis, findings and save 390 human-hours per year.
EXIM End to End Automation:
• Crafted a Power Automate pipeline to store data into Azure Blob, pulling the same data in Azure
Databricks.
• Utilized,cleaned & transformed raw data in Azure Databricks using python spark framework with 80 %
more efficient by optimization.
• Stored the cleaned data in SharePoint to pull cleaned data in Qlik for identifying purchase trends, market
patterns and competitor pricing thereby empowering decision making.
• Managed time saving of 400 human-hour per year through this ETL automation and streamline the
process.
Cycle Time Analytics:
• Engineered a Qlik Dashboard for Manupur Plant to analyze time for API formulation which lead to increase in
efficiency by 30% and contributed to cost saving of approximately $130 per month per block for FY 23 with
this streamlined and low level latency dashboard solution
Associate Jan 2020 - Apr 2020
•
Altudo Gurgaon, India
Analyst:
• Translated complex product requirements into precise analytical specifications, resulting in a 20% increase in
development efficiency.
• Successfully loaded and transformed data from PostgreSQL, enabling seamless integration with machine learning
models, saving 15 hours of manual data preparation time per week.
• Spearheaded data preprocessing techniques that significantly improved the accuracy of machine learning models
by 25% and reduced potential errors in predictions by 30%.
Personal Projects
• APS Fault Detection: Github
• This project aims at classifying whether a failure cause in heavy duty vehicle is caused by APS system or not . Where
APS stands for Air Pressure System. An End-to-End project where affirmative class means failure is due to APS
and negative class means Non APS system.
Education
St.Francis Institute of Technology Mumbai, India
•
Bachelor in Electronics and Telecommunication Aug 2017 - Jul 2021