Customized Problem Solving with AI

We have created and delivered a wide range of customized solutions with consultative services for specific customer problems. Most feasible AI technologies and software architectures are selected case by case depending on the problem area and customer needs. 

Price Forecasting Solution

Price Forecasting solution to forecast future prices of products of specific industry domain. Forecasting is done based on tens of different data sources of various data types.

The solution included an easy-to-use web-application to modify the AI / ML based forecasting according to market dynamics and business priorities. Various filtering and theme selection functionalities were created to view the results for different purposes.

Recipe Optimization with AI

The solution included application functionalities for data processing, training machine learning models, running predictions and deriving optimized recommendations.

Up-scaling of the working processes was enabled through cloud-based application architecture and solution deployment.

Cash Flow Prediction

The project included creating a deep neural network based solution to predict end of year cash flow to enable more accurate financial planning and forecasting.

Model training was made using millions of rows of sensitive invoice data.

 

Industrial Instance Segmentation

Implementation of a computer vision solution to perform industrial instance segmentation and counting. 

The solution enabled aggregated statistics of the instance sizes & shapes and intuitive visualization of the prediction results.

Fiber Counting solution

The solution contains automized quality inspection of microscopic product samples using AI / ML technologies.

Solution included an intuitive web based applications for laboratory personnel that enabled streamlining the work procedures.

AI for Smart Device Test Robot

The solution included a deep learning based cosmetic fault detection engine for a smart device test robot.

The solution achieved high accuracy and flexibility for different screen quality classification fault categories. The result was achieved by using only small data set of good/defective items.

Want to hear more?

Customized Problem Solving with AI

We have created and delivered a wide range of customized solutions with consultative services for specific customer problems. Most feasible AI technologies and software architectures are selected case by case depending on the problem area and customer needs. 

Price Forecasting Solution

Price Forecasting solution to forecast future prices of products of specific industry domain. Forecasting is done based on tens of different data sources of various data types.

The solution included an easy-to-use web-application to modify the AI / ML based forecasting according to market dynamics and business priorities. Various filtering and theme selection functionalities were created to view the results for different purposes.

Recipe Optimization with AI

The solution included application functionalities for data processing, training machine learning models, running predictions and deriving optimized recommendations.

Up-scaling of the working processes was enabled through cloud-based application architecture and solution deployment.

Cash Flow Prediction

The project included creating a deep neural network based solution to predict end of year cash flow to enable more accurate financial planning and forecasting.

Model training was made using millions of rows of sensitive invoice data.

 

Industrial Instance Segmentation

Implementation of a computer vision solution to perform industrial instance segmentation and counting. 

The solution enabled aggregated statistics of the instance sizes & shapes and intuitive visualization of the prediction results.

Fiber Counting solution

The solution contains automized quality inspection of microscopic product samples using AI / ML technologies.

Solution included an intuitive web based applications for laboratory personnel that enabled streamlining the work procedures.

AI for Smart Device Test Robot

The solution included a deep learning based cosmetic fault detection engine for a smart device test robot.

The solution achieved high accuracy and flexibility for different screen quality classification fault categories. The result was achieved by using only small data set of good/defective items.

Want to hear more?