Machine Learning
Sumth has a team of skilled ML architects, data scientists and machine learning engineers that can formulate the right ML problem definitions, define the right metrics and loss functions, extract, clean and validate data, build models, deploy them, and continuously improve the models to achieve desired metrics..
We have experience in variety of classical and deep ML models for use-cases around marketing, next-best offers, churn prediction, campaign optimization, predictive maintenance, revenue forecasting, demand forecasting, inventory optimization, logistics, planning and scheduling, credit scoring, price & demand sensitivity models, fraud detection, etc.
Improve Campaign Efficiency
Achieved a 10x improvement in campaign efficiency over hard-coded offers for a telecommunications company using an NBO framework consisting of multiple propensity, persuadability and revenue optimization models.
Increased Take Up Rates
A 400% improvement in take-up rates was experienced, for customised offers using active learning, hierarchical clustering and multi-armed bandit optimization to come up with the right campaign strategy.
Reduced Errors
Estimated reduction of 90% in passenger misconnections in over 3 months using .a multi-objective optimization system in one of Asia’s busiest airports
