Now you can Instantly Chat with Arjun!
About Me
A certified Data Scientist with 5 years of software and research experience in predictive analytics to provide solutions for the business by generating data driven insights. Capable of presenting insightful and actionable business recommendations ...
Show MoreSkills
Portfolio Projects
Description
Problem Statement: Apply Machine Learning models to evaluate the price based on Mileage and Grade for used vehicles to project on online auction sale platform using RapidMiner web-services and Model Deployment.
• Methodology: #VIN details are captured to apply Deep Learning Model(H20), Gradient Boosted Trees and Random Forest to identify outliers, predict demand label and apply business logic to define Price and Click&Buy Price for online auction sale platform.
• Responsibilities: Model Development, Integration, API Development, Test, Documentation,
Show More Show LessDescription
Problem Statement: Understanding customer dynamics and behaviour in ecommerce settings
• Methodology: Customer Segmentation using RFM Analysis and Market Basket Analysis using R Studio
• Responsibilities: Quantifying RFM values, clustering to discover groups and classification to differentiate and predict. •
Show More Show LessDescription
Problem Statement: Identify appliances from a home tour video to estimate the price of the home insurance
. • Methodology: AWS Rekognition API and TensorFlow Object Detection API to quickly visual search, classify, detect objects and label.
• Responsibilities: Integration, test, documentation Python SDK - AWS Rekognition and Keras
Show More Show LessDescription
Product review, extensions development and test new features.
• Provided Solution and technical support for the issues/tickets raised by clients as per SLAs
. • Handled customer calls and demos of RapidMiner Software for Pre and Post-Sales requirements.
• Lead a team of 5 data scientists to troubleshoot the issues of RapidMiner
Show More Show LessDescription
Problem Statement: Proactively prevent fraudulent incidents by monitoring payment transactions info.
• Methodology: Developed advanced self-learning rules engine that has mechanisms of soft thresholds for establishing and interpreting “normals” using AI-based platform (developed using RapidMiner)
• Responsibilities: Implementation of multiple techniques including rule-based, supervised and unsupervised machine learning to understand, predict and act in real time to detect fraud.
Show More Show LessDescription
Responsible for providing statistical support for Systems Biology lab (Data extraction, data validation, cleaning and investigation of missing values using R)
• Performed clustering analysis on various cancer datasets (Breast, colon, kidney, lung and prostate) to estimate groups for further research.
• Provided statistical analysis using R and SAS for analysis of variance, hypothesis testing (t-test, chisquare).
Show More Show LessDescription
Designed, programmed, debugged installed and maintained department’s database (MS Access 2007)
• Added automated functions to existing access databases and redesigned the queries
• Gave outreach presentations in various public schools to educate students about science and engineering.
• STEM Student representative for Padnos College of Engineering - GVSU 2009.
• Provided technical support for installation and troubleshooting services to hardware, software and peripheral equipment, following design and installation specifications.
• Provided technical support in resolving queries of students and staff related computer programs and performance related problems.
Show More Show LessDescription
Problem Statement: Marketing performance measurement and improve current process and systems for effective marketing insights delivery.
• Methodology: Marketing measurement techniques such as multi-touch attribution, Marketing Mix Model, segmentation and lifetime value model development.
• Responsibilities: Lead and team of 3 Data Scientists for applying advanced data analysis (regression, key drivers, segmentation, LTV modeling) to translate insights into implications for marketing.
Show More Show LessDescription
Problem Statement: Lead scoring and decision tree analysis to identify key lead segments and score at each stage of the business
• Methodology: Lead segmentation, Scoring and Model recommendations using Rapid Miner
• Responsibilities: Transformation, cleansing and preparation of data from Amazon RedShift DW (70,000 rows) in suitable format for scalable analytics. Feature extraction and implementation of machine learning algorithms to achieve best possible results from our predictive models. Tableau dashboards for validation of scores from predictive modelling.
• Tools: Rapid Miner, R Studio, Salesforce, Marketo, Tableau and SQL-Workbench
Show More Show Less