Inventory Optimization

Assembly Line

How Instacart Uses Machine Learning to Suggest Replacements for Out-of-Stock Products

📅 Date:

🔖 Topics: Recommender Systems, Inventory Optimization

🏢 Organizations: Instacart


One of Instacart’s key challenges is predicting product availability without real-time inventory data. Our machine-learning model prompts replacement suggestions if a product appears unavailable when an Instacart customer shops. This replacement model also assists Instacart shoppers in selecting the best replacements during their shopping trips.

Our model uses a Siamese network that leverages identical weights to simultaneously process two different input vectors, creating output that can be easily compared. This configuration mirrors the classic ‘two-tower’ architecture prevalent in recommendation and search ranking applications. The product layer consolidates the four types of features mentioned above into an embedding representation for a product. The model employs a BERT-based sentence embedding layer to process product name text features, and embedded representations for high-cardinality categorical features are learned from scratch during model training.

Read more at Instacart Tech Blog

Inventory allocation optimization: A pre-built solution for Dataiku

📅 Date:

✍️ Authors: Marko Nikolic, Andriy Drebot, Ilya Katsov, Dejan Dzunja

🔖 Topics: Inventory Optimization

🏢 Organizations: Grid Dynamics, Dataiku


In this blog post, we present the Inventory Allocation Optimization Starter Kit for complex environments that was jointly developed by Grid Dynamics and Dataiku. This solution is implemented on top of the Dataiku platform and is available in the Dataiku marketplace. It is geared toward retailers, brands, and direct-to-consumer manufacturers that seek to improve supply chain operational efficiency, increase customer satisfaction, and reduce shipping costs and order splits.

We show how you can make more informed inventory level decisions to meet customer demands and cut costs by carefully considering a variety of factors, including demand forecasting, procurement, order splitting, and shipping costs with a single solution: The Inventory Allocation Optimization Starter Kit. The solution can be extended with various features such as procurement shipping capacity constraints, region-level constraints, and more. The created pipeline can also be transformed into a replenishment control solution that optimizes not only the static inventory allocation levels, but also replenishment and shipping times.

Read more at Grid Dynamics Blog

How Walmart Uses Apache Kafka for Real-Time Replenishment at Scale

📅 Date:

🔖 Topics: Inventory Optimization, Demand Planning

🏢 Organizations: Walmart, Confluent


Real-time inventory planning has become a must for Walmart in the face of rapidly changing buyer behaviors and expectations. But real-time inventory is only half of the equation. The other half is real-time replenishment, which at a high level, we define as the way we can fulfill the inventory demand at every physical node in the supply chain network. As soon as inventory gets below a certain threshold, and based on many other supply chain parameters like sales forecast, safety stock, current availability of the item at node and its parents, we need to automatically replenish that item in a way that optimizes resources and increases customer satisfaction.

On any given day, Walmart’s real-time replenishment system processes more than tens of billions of messages from close to 100 million SKUs in less than three hours. We leverage an array of processors to generate an order plan for the entire network of Walmart stores with great accuracy and at high throughputs of 85GB messages/min. While doing so, it also ensures there is no data loss through event tracking and necessary replays and retries.

Read more at Confluent Blog