Price Optimization for SaaS Product

The founding team wanted a first fully-scalable product from what they had envisioned from day 1. They wanted a product that would fetch different data sources, apply several machine learning models to them, optimize prices across different goals and constraints, and push prices back to the e-commerce platform.

Built an end-to-end MLOps pipeline for dynamic pricing optimization using quantile regression as baseline and Lagrangian methods. The system processes e-commerce data through automated workflows, trains demand curve models, and applies multi-objective optimization strategies (profit maximization, sales growth, constraint-based) to generate optimal pricing.

The system was able to generate optimal pricing for the e-commerce platform, with a 35% increase in revenue and a 25% increase in sales.

Taking the founding team's vision and early design to a fully data-intensive product and iterating over customer feedback unlocked new revenue streams for the company.