Oct 2020 – Present · 5 yrs 7 mos · On-site

• Currently working in the Supply Chain Optimization Technologies (SCOT). Work includes solving very-large scale network flow problems for generating short-term plans for Amazon's transportation network, and using Machine Learning for generating demand inputs for the planning systems.
• Built a Machine Learning based system for predicting which shipments going through Amazon Air are at a risk of not meeting customer promise, with customized precision/recall score of more than 80%.

Oct 2015 – Sep 2020 · 4 yrs 11 mos · On-site

• Built systems for Amazon Air network planning from scratch, including multiple Optimization/ML models that enabled the growth of the business from zero to 100+ aircraft in less than six years. Enabled faster delivery of billions of shipments for Prime customers over the years, resulting in multi-hundred-million-dollar transportation cost savings for Amazon.
• Developed a Machine Learning model for generating delivery windows for shipments
• delivered by Amazon Logistics, that has more than 95% accuracy and generates predictions for
• more than 90% of the shipments. Available as a customer facing feature in Amazon shopping
• app in US and EU since 2019, and it has greatly improved delivery experience for Amazon
• customers.

Sep 2011 – Sep 2015 · 4 yrs · On-site

• Built analytical tools for increasing the footprint of Amazon’s last mile delivery network in multiple major markets.
• Helped grow the UK last mile network from ground up to delivering more than 80% of Amazon’s shipment in UK within four years, resulting in faster shipment delivery and millions of dollars in cost savings.
• Devloped models for expanding Amazon Logistics network in multiple other countries: US, India, Germany, Italy, Japan.