Shubham Devendra Saxena

After completion of my Trainee program, I started working as a Data Scientist.

• Created ML pipeline which pre-processed data with inclusive usage of isolated forest ML model to predict potential customers. The ML model successfully predicted customers who signed contracts amounting to half a million euros.
• Enhanced data-driven marketing decisions making by implementing machine learning models and algorithms
• Implemented hyperparameter tuning for LightGBM to improve its performance for a classification problem
• Collaborated with business-unit to understand specific business requirements and translating them to data science use-cases
• Optimized machine learning pipelines and computational resources deployment strategies resulting in reduced processing times
• Built Python code base which sends HTTP request to data processing services and returns processed data
• Dived deeper into usage of Azure ML for running jobs on Azure compute clusters
• Implementation of codebase based on Python OOP
• Arranged code reviews for faster and high-quality development of code

The trainee program was for one year, wherein I underwent rotation across three departments

Phase 1- Data Science and Analytics

• Analyzed large datasets to identify trends and patterns in customer behaviors
• Generated insights to drive marketing decisions
• Developed polished visualizations to share results of data analyses
• Mapping of business area requirements to data science use cases
• Learnt implementation of best practices like unit tests in Python code and modular structure

Phase 2- MLOps

• Deepened knowledge into using MS Azure
• Became proficient in Azure DevOPs, emphasizing project management features
• Created a REST API add-on to an existing CI/CD pipeline in Azure DevOps to interact with a service
• Created a front-end for a chatbot using Django

Phase 3- Data Engineering

• Deeply engaged with Azure Data Lake Storage (ADLS)
• Created a front-end using Django to interact with Azure Blob storage for data upload to the data lake

• Created classification machine learning model based on XGBoost and RandomForest
• Integrated new machine learning models in existing code pipeline
• Normalization of time series data using pandas.

• Performed resistance spot welding tests for data acquisition
• Root cause analysis to eliminate anomalies in data
• Created XGBoost regression model for predicting spot-welding nugget diameter
• Model performance visualized using PowerBI.

• Data Analysis of the sales data provided by Grünenthal GmbH
• Visualized the product sales using Python
• Wrote code in Python to handle the large amount of data and plotting numerous plots for analysis
• Finally determined the effect of a launch of a new product variant on older product variants

• Sensor Signal data evaluation and visualization in Python
• Adjustment and calibration of spectrometers
• Automated generation of test protocols.