Transform Reactor Data

Critical Decision-Making

AI Driven

Batch reactor
process prediction

At Neural Flow, we use Machine Learning and Artificial Intelligence (AI) technologies to turn data collected in any standard manufacturing process into a powerful decision-making engine for optimizing processes, maximizing yields, and ensuring consistent product quality.

With Neural Flow, the information you need in order to make informative, calculated decisions is at your disposal in real-time, all the time.

Real time data analysis that allows immediate response
Identification of causes for suboptimal batches
Prediction of batch output
Keeping knowledge in the company
how it works
Real-time analysis that allows immediate response

As process engineers, the decisions we make rely heavily on data, and the more of it, the better. Tools based on BI and data analysis used today may seem sufficient, but the information they gather and process is limited, and so is the analysis they provide. We still need to do most of the job “manually”.

Neural Flow changes this dynamic

Neural Flow’s system automates data collection, adding a wide range of information that is there but never used, and analyzes it based on benchmarks and historical data to predict the outcome throughout the process. With real time information, process engineers are able to make necessary adjustments throughout the manufacturing process and guarantee an optimal final outcome.

Using machine learning and AI, the system examines correlation between sensors, compares current batch to historical ones, and identifies non-linear correlations. Analysis is presented clear and simply on the Neural Flow dashboard, indicating whether the production line is optimal or not.

With a system that allows to predict outcomes in real time, companies are able to prevent process inefficiencies, continuously improve production capabilities, and cut costs associated with suboptimal production.

In addition to the dashboard, real-time alerts are sent to designated users, allowing them to immediately react to any deviation and manage suboptimal batches, before it’s too late.

learn more

A ook into Neural Flow’s machine learning engines

The prediction capabilities of Neural Flow’s system are achieved thanks to three machine learning engines; each has its unique role in the process.

Neural Flow uses three types of machine learning engines.

Clustering - Detects anomalies in batches

learn more

Neural Network - enables time series data prediction.

learn more

Statistical -signal analysis & forecasting

learn more
Autonomous identification of causes for suboptimal batches

Without proper tools, identifying the root cause of reactor malfunction is a complex and time- consuming task. BI, data-analytics, excel and others technologies available today, provide information, but the analysis and conclusions still need to be done manually.

 Neural Flow offers a new approach

Neural Flow’s system uses mathematical tools to remodel batch analysis, identifying possible causes for deviation in quality. Analysis is presented in real time as actionable data, complemented by linear and non-linear sensor correlations and comparison to historical data, allowing process engineers to easily identify visible and hidden reasons for batch under-performance.

In addition to real-time monitoring, the Neural Flow platform also enables exploration of historical processing problems. It allows to drill down and review different parameters of each batch, and identifies trends and behaviors using state-of-the-art data analytics tools.  Information is presented in our application via easy-to-read pre-defined dashboards; customization options also include the use of graphs and comparison tools.

By autonomically identifying the root cause of suboptimal batches, the Neural Flow system saves time, while enabling a more accurate and efficient process with optimal control on results. 

learn more
Prediction of batch output, during production

Predicting the outcome of a batch, during the process, allows to make necessary changes as we go. It lets us correct and align reactor parameters, and change the outcome before it is too late.

Neural Flow opens a window of opportunity

The Neural Flow system uses data provided by two independent sources: reactor sensors, and lab results. Our proprietary algorithms perform calculations based on data from both sources and synchronizes between them to achieve maximum accuracy.

Accurately predicting the outcome of a process gives process engineers an opportunity to correct and align reactor parameters during the process, limiting the occurrence of faulty batches, reducing costs and saving valuable time.

learn more
Keeps knowledge in the company

Managing knowledge and making sure it remains within the company is the best way to avoid brain drain which unfortunately occurs when an engineer leaves for another company.

Neural Flow offers a solution

By automating data collection and its analysis, and managing historical data in a secure platform, Neural Flow helps limit the dependency on specific individuals.

The Neural Flow platform is highly secure and can be implemented according to local regulatory requirements and IT demands. The system can either operate on a secure and encrypted cloud or be limited to an on-premise server where the data cannot leave the site.

Neural Flow guarantees that knowledge remains safe and secure in the company, and continues to serve it in decision-making.

learn more

The Neural Flow platform is highly scalable. It is a semi-generic platform that offers you the best of both worlds. On the one hand it fully customizes itself to the characteristics of your production facility, and on the other hand it uses easy to implement software. The scalability process is very fast and cost-effective, and the system can be easily applied to many processes within a single manufacturing facility.


The Neural Flow platform is highly secure and can be implemented according to local regulatory requirements and IT demands. The system can either operate on a secure and encrypted cloud or be limited to an on-premise server where the data cannot leave the site.


Founded in 2019, Neural Flow is an Israeli-based 4.0 startup and part of the MGT Liquid & Process Systems Group, an industry leader with over 50 years of experience in manufacturing custom reactors, mixing systems and process solutions.

As process and chemical engineers and mathematicians, our mission is to streamline process management, utilizing information that is usually overlooked in the standard way of doing things. Our software platform is designed to assist engineers in their daily work, and provide tools to improve process results and company profitability.

The Neural Flow system is implemented in the chemical, pharmaceutical, paint & ink, cosmetics and F&B industries, and is suitable for any process involving reactors.

We look forward to demonstrating our system’s capabilities and show you how we can help you achieve better results.