I am part of cloud supply and chain team. Currently, I am looking after new demand forecasting techniques.

Data and Applied Scientist II

Handling In_Store campaigns from planning to execution, targeting to evaluation
Generating high uplift for suppliers

Lead Applied Data Scientist

Forecasting and prediction :- Sales forecasting was done for more than 200k parts using various different time series techniques i.e. ARIMA , Holt - Winters , BATS, TBATS etc. -- Results were compared with legacy methodologies and an improvement of 18% was observed resulting in $40M of estimated annual operational cost saving

Lead Data Engineer

NTMS/CRM/Siebel management, Data Visualization, Machine Learning, Churn Model, Segmentation model  

Business Analyst

1.Python is the primary technology I have worked on, I have done it all from doing data manipulation to using advanced neural networks
2.Data Driven Business Operations: Successfully adhered to the TRAI and company data governance by removing duplication using fuzzy logic and then masking the confidential attributes of customers using encryption script in Python



1.Customer Segmentation Model: Segmentation of customers was done using K-Means Clustering algorithm in python during 4G Launch to maximize penetration and for better business of newly launched services. --Model acted as a deciding factor to provide 4G services to maximum possible subscribers with in 90% of available capacity

2. Customer Churn Prediction Model- Built a machine learning classification model using machine learning using many classification techniques like Logistic Regression, Random Forest and Boosting techniques to calculate the probability of each customers from the network using the past data and attributes of more than 3 million customers. --The overall precision of model was ~95% and churn rate was reduced from 2% to 0.95%. 


2.Created several simple as well as complex machine learning models like Linear Regression, Logistic Regression, NLP, PCA. Neural Network, SVM etc

3.Time series forecasting models : Arima, Auto Arima, bats, tbats, prophet, ets etc,

4. Natural Language Processing: LDA, TF-IDF, Bag of words, N-gram


1.Dash-boarding and Reporting- Created several Tableau reports to track real time circle wise sales ,customers data & other KPIs and automated recurring reports to reduce active time by 90%

2.Fraud and Leakage Revenue Analysis - Identified international roaming partners using past data to balance out packet units, minimizing the transactions and saving taxes. --After the analysis there was a leakage of $30000 per month was detected

1.Device Repository Management- Creation, Development, up-gradation of device repository in SQL environment for PAN INDIA and update new TACs into the ecosystem of device management for business intelligence, IT queries, promotions and CRM improvement. --Generated and verified reports of demographically band wise distribution for PAN INDIA for the SPECTRUM bidding analysis.





1. Hypothesis testings on several projects, 
2. Uplift calculation through ANOVA, ratio method
3. Several brand analysis for a campaigns



Personal details

Aman K.

Summary

Technical skills

Work Experience

Education