Data Consumption - Predictive Modelling & Analytics
Customer 360 Model
Built & Deployed a customer360 model, serving as a comprehensive system for managing customer data.
Key features are the integrated ML model scores such as propensity, abnormality, and Uplift for various interests and determine the best offers for each customer based on their distinct traits and behavioral patterns.
Up-Sell & Cross-Sell
Expertise in building & deploying predictive models utilizing various ML algorithms to identify targeted customers likely to accept Up-Sell or Cross-Sell offers.
Key takeaways are the understanding of the core customer priorities & special needs when recommending an offer, Careful selection of features required for the optimal targeted customer.
Credit Risk Model
Designed & built a model that predicts the Probability at default time where customer could not repay the money. Also the ML Model predicts the Exposure during the default instance and the % of loss from the overall loan credits dispersed.
The key predictors are the customer income range, credit agency ratings, house Ownerships, decile, higher education, and social & environmental factors.
Operational Efficiency
Designed & guided the team to build model that can forecast the completion of batch pipelines. The primary predictors are current CPU resource availability, CAP & Utilization, database, and network loads. With 1000s of jobs running achieving operational excellence is a monumental challenge, ML model helps to know the information of lead time to complete so that the operations team plan their work well saving their efforts.
Predictive Modelling for Churn Reduction
Expertise in building & deploying Predictive models for churn reduction using boosting algorithms to identify potential churners. Key insights include effective feature Selection can result in at least a 15-30% improvement. Incorporated features related with the customer status, Credits availed & usage behavior.
Customer Life Time Value
Designed & built a model that predicts the Customer Life Time Value for the future.
This need prediction as this may need to consider the future variability factors including factors that affect the revenue, gross margin, and churn %
The Key Predictors are the loyalty stickiness, Price Sensitivity of the customer, History of number of times churns, change in the products, and trouble tickets counts.
AI Governance & MLOPS
As part of AI Teams, Ensured ML OPS standards for new deployments, Documented Data Governance standards & best practices, KPIs, success meetings & communications for stakeholders.
KEY PERFORMANCE INDICATORS FOR DECISION SUPPORT
Spearheading the product team on metrics calculations, ensuring the correctness, conducting workshops to clarify and confirm with the business teams.
​
GOLDEN DATASET CREATION - BALANCING DATASETS
Involved in building the Activation functions for an ML Models so that the problem statement can be arrived and golden datasets can be built for the interest based scores and traits based measurements
Generative AI
Practical Implementation of a Simple Lang chain Solution — A Step-by-Step Guide - Introduction
Data Science - Decision Boundary of Supervised Learning Al
How Generative AI Models inspired by core math, stat, and ML concepts
