I led the technical design and implementation of an ML-based delivery time prediction system that modeled the client’s supply chain end-to-end. We built an ultra-low latency service leveraging Rust and low-level optimizations as well as a high-throughput precomputation pipeline that is precomputing delivery promises for Wayfair’s catalogue of 10M+ products every few hours.
The system powers delivery promises across massively impactful use cases such as the delivery promises on the search, browse and product page experiences as well as feeding external systems such as Google’s product feed.
Interested in a similar solution? Let's talk.