Customer
Decision acceptance deadline

01.12.25 (inclusive)

Form of award

Under discussion

Product status

Idea

Task type

ICT tasks

Сфера применения

Food industry

Область задачи

Other technological solutions

Type of product

Software/ IS

Problem description

Currently, demand planning is done manually and is based on the subjective assessment of managers. This leads to underloading of couriers during quiet hours and overloads during peak periods. As a result, delivery time increases and costs increase. We need an automated solution that provides a stable and accurate forecast.

Expected effect

Increase the accuracy of demand forecasting by 20-30%, optimize the distribution of couriers, reduce transaction costs, reduce delivery time and improve the customer experience.

Full name of responsible person

Temirbulatov Bekbolat Bauyrzhanovich

Purpose and description of task (project)

Creating a machine learning-based solution for accurately predicting the number of orders by area and time interval. The goal is to automate demand forecasting in order to reduce transaction costs, optimize the number of couriers, and improve customer service. It is necessary to develop a model that automatically analyzes historical orders, time patterns, weather, holidays and local activity, forming an accurate load forecast for the day and week.

Note

We are ready to provide additional information about the task and technical requirements upon request of moderation.