Introducing world's first Potential Mining Platform

Connect your data and use cutting edge AI powered Predictions and Multi-Objective Optimization – easy to connect with only a few lines of code or even no code – provided as a Service without any implementation costs! 

paretos offers the power of an outstanding algorithm “socrates” to deal with high-dimensional, complex models – easy to integrate in existing tools and with comprehensive data analytics

Manage Complexity

Use the potential of the paretos dynamic Prediction and Optimization Service to power recommendation, matching, pricing engines or other dynamic topics with changing data and parameters. The data or a model – if available – can be connected to the paretos platform in less than two hours and off you go!

Use ML at best quality

Use and optimize Machine Learning Algorithms & Neural Nets in Application – fully automated! You can either use paretos Prediction Models (without any ML knowledge) or connect your own models and algorithms with paretos’ A.I. based Optimization Engine.

Predict your full potential

Forget manual iterations! Visualize all predicted scenarios based on your data. Find customized, pareto-optimal designs and scalable operational strategies with paretos’ A.I. based Optimization Engine. The paretos API can be easily connected to all state-of-the-art tools and systems.

Our Partners

Why paretos?

Predict

the full technological potential based on your data with paretos Predictions as a Service or connect an own model – via API and without any integration costs

Optimize

the potential of your Data or your ML Models with paretos dynamic Multi-Objective-Optimization algorithm “Socrates“, integrated smoothly via our API

Derive

strategic decisions by visualizing the results and use next level predictive analytics – with an interactive dashboard… potential mining “as a service”

Predictive Analytics

Predictive Analytics

Success Stories

paretos optimization algorithm Socrates has been validated in complex projects, such as the automated development of neural nets for bee recognition, development of an Electric Vehicle or the design of hybrid powertrains. They serve as a showcase of the significant increase in efficiency.

System time reduction
through integrated automation

75%

less time

Simulation Time
With full factorial observation

5h

instead of 84 weeks

Neural net performance boost
for fast and precise bee recognition
through efficient optimization

99%

recognition rate

Training time
compared to full factorial evaluation
where each training run takes up to several days

167

instead of 168,480 runs

Potential unlocked
by showing best train alternatives for a specific track
through multi objective optimization engine Socrates

18

optima revealed

Simulation Time
Compareted to full factorial observation
where one train simulation takes one minute

1h

instead of 27 weeks

What others say

FAQ

paretos provides Predictions and Optimization as a Service. By showing the full potential of your data, you can dynamically derive decisions at any time – even in rapidly changing environments with high complexity.

You can enter the platform either with connecting / importing your data (via your ERP, CRM or simply via .csv (or Excel) import) or connecting your model directly.

You can use paretos Potential Mining, Predictions and Optimization for every complex question in dynamic environments. Our customer use it for re-inforcement learning within recommendation engines, dynamic matching algorithms, inventory&supply chain predictions and optimization, predictive analytics and pareto-optimal simulations (with competing objectives)

paretos AI-based Optimization Engine can be integrated quickly, easily and securely via an API, while the raw and model data is kept on the customer’s server. 

By a dockerized model layer and a lean meta data interface on the client side, paretos is able to provide the AI-based optimization via an API. Finally, the paretos optimization engine triggers the model optimization fully automatically and highly efficiently, receives all optimal results of a technology potential of a simulation or a Hyperparameter Tuning for Machine Learning Models.
In addition we provide data analytics and visualizations based on the learning of the model optimization.

Beside the fact, that with paretos you are able to integrate AI Optimization to your business as easy as to connect to a new mail-service, our Optimization Algorithm was researched over several years at the University of Applied Sciences Munich and has been recognized and validated worldwide (e.g. by the MIT). It is now being further developed and transferred to the application by paretos and outperforms state-of-the-art optimization Algorithms in high-dimensional spaces. Feel free to get in touch with us for mindblowing benchmarks! 

You can use paretos Potential Mining, Predictions and Optimization for every complex question in dynamic environments. Our customer use it for re-inforcement learning within recommendation engines, dynamic matching algorithms, inventory&supply chain predictions and optimization, predictive analytics and pareto-optimal simulations (with competing objectives)

for connecting models directly: our optimization engine works with every model and model management tool. It can be used for Machine Learning models, design (hyperparameter tuning) of neural nets (e.g. for prediction or object recognition) or for virtual prototypes, Digital Twins and other complex simulations. 

But beside of that: Get in touch with us today to find out how we can integrate AI based Optimization in your company or for further use cases and success stories!

You can enter the platform either with connecting / importing your data (via your ERP, CRM or simply via .csv (or Excel) import) or connecting your model directly:
You can connect the model via API calls (just a few lines of code – see the doc here) or via all common model management tools (such as TensorFlow, PyTorch, SageMaker, Matlab, Modelica). The raw and model data is kept on the customer’s server. Our optimization engine only requires the hyper parameters, the meta performance and the optimization metrics. By a dockerized model layer and a lean meta data interface our Optimization Engine triggers the model optimization fully automatically and highly efficiently.

As we provide our Optimization as a Service (per API calls) we charge our clients by a monthly subscription fee or a “pay per use” fee. The set-up costs and effort is very low, as we can communicate with all common tools and languages (such as TensorFlow, PyTorch, SageMaker, Matlab, Modelica)

Leadership

Fabian Rang

Founder

Fabian is fascinated by the potential mathematical optimizations for the design of complex problems. Raised as an engineer at a huge german OEM, he went into Systems Engineering and maths – his PhD is about “Efficient multi objective optimization for expensive high-dimensional blackbox problems”. He co-founded paretos with the urge of closing the gap between maths and business applications.

LINKEDIN // YOUTUBE

Thorsten Heilig

Founder

Thorsten is fascinated by the potential mathematical optimization can bring to complex (real-life) problems. With focus on organizational & product growth, he co-founded companies, worked as a management coach and recently was COO with moovel/REACH NOW before co-founding paretos. His motivation is to provide AI-based technology for “everyone” to make better strategic decisions. 

LINKEDIN // TWITTER 

Contact

paretos "in a nutshell"

paretos provides Potential Mining, Predictions and Optimization – as a Service.
By showing the full potential of your data, you can dynamically derive decisions at any time – even in rapidly changing environments with high complexity.

Our customer use it for re-inforcement learning within recommendation engines, dynamic matching algorithms, inventory&supply chain predictions and optimization, predictive analytics and pareto-optimal simulations (with competing objectives)

You can enter the platform either with connecting / importing your data (via your ERP, CRM or simply via .csv (or Excel) import) or connecting your model directly. Thus paretos enables companies to use state-of-the-art optimization methods in development and decision-making processes. The software and the optimization algorithm can be easily integrated into the existing IT landscape, raising new technology potentials and leading to dramatic efficiencies.

The AI-based optimization algorithm was researched over several years at the University of Applied Sciences Munich and has been recognized and validated worldwide (e.g. by MIT). It is now being further developed and transferred to the application by paretos.