Toptal improves model accuracy by 2x and data speed by 50x for footfall analytics firm.

Toptal partnered with a footfall analytics firm to upgrade its outdated algorithms, improving the accuracy of its predictions and the overall quality of its service while addressing evolving Wi-Fi signal patterns.

Client

A Sweden-based robotics company developing autonomous pick-and-carry robots for industrial and logistics use.

Employees

20+

Revenue

$3M

Industry

Technology

Delivered Services

Get a free consultation now

Schedule a Call

Challenge

The firm’s outdated algorithms struggled to adapt to evolving Wi-Fi signal patterns, leading to a decline in the accuracy of footfall analytics and threatening the firm’s reputation for reliable data.

Solution

Ground Truth Data Collection

Toptal developed a data-driven strategy to enhance the collection of ground truth data, enabling more accurate validation and calibration of the firm’s footfall analytics models.

Machine Learning Framework

Toptal established a robust framework for conducting and documenting machine learning research, incorporating SQL databases and Python, to ensure that the firm could adapt its models to future technological changes.

Outcome

2x Accuracy, 50x Speed Improvement

Algorithm enhancements reduced the out-of-sample model error by 2x and accelerated data processing by over 50x, boosting the accuracy and efficiency of the firm’s footfall analytics.

24/7 Data Pipeline

Toptal implemented a 24/7 data pipeline using AMQP, enabling continuous and efficient data collection and processing, which laid a foundation for ongoing machine learning research and future improvements.

Share

Get a free consultation now

Schedule a Call