Modelcraft | Management Solutions

Why ModelCraft™?

ModelCraft™ is an automated machine learning and component- based modelling tool designed to simplify and industrialize the development and validation of advanced analytics models

ModelCraft™ has intuitive components for data processing, advanced analytics model development and interpretability. This platform empowers users to seamlessly connect to any data source and create modeling workflows visually, eliminating the need for coding expertise.

ModelCraft™ effectively meets industry demands by:

Simplifying the integration of sophisticated machine learning algorithms—such as gradient boosting, elastic nets, ensemble models, random forests, support vector machines, and neural networks—into the modeling process.
Streamlining and expediting the modeling process through code reuse and minimizing the time dedicated to routine tasks.
Making advanced modeling techniques accessible to teams without specialized knowledge.
Enhancing the performance, traceability, and comparability of models.
Facilitating model interpretability and automating the generation of model documentation.
Ultimately, refining model development and validation processes and fostering the creation of advanced challenger models.

ModelCraft™ is developed in the cloud and natively uses auto-scalability and parallel computing capabilities.

In addition, the ModelCraft™ ecosystem includes the ModelCraft™ Library suite, a Python/PySpark library built on reusable analytical components that are tested during development. This enables the structured, standardized, and accelerated development, validation, and deployment of models through robust, reproducible, and auditable pipelines.

Thanks to this, ModelCraft™ tool accelerates the transition to component-based modelling, which is more efficient, interpretable, and accessible, while reducing model risk.

Functionalities of ModelCraft™

ModelCraft™ features a library of predefined components that support the entire modeling lifecycle

Input, connection and download

Input data reader supporting different formats and delimiters

Database connection components (Cloud, RDS)

Download of models, processed datasets and automated reports

Sampling

Sampling algorithms

Training, test and validation samples

Algorithms for processing highly unbalanced data (oversampling and undersampling techniques)

Visualizations

Model comparison and selection

Analysis of database structure

Exploratory Data Analysis

Calculation of descriptive statistics

Analysis of database structure

Calculation of descriptive statistics

Performance and prediction

Mapping of criteria to training samples, validation and testing

Calculation of metrics and KPIs

Out-of-time analysis and backtesting

Predictions

Model validation

Database management

Simultaneous connection to different data repositories, including MySQL databases

SQL data processing (generation of tables, data manipulation and retrieval, table traversal, etc.)

Import and export files

Feature engineering

Correlation analysis

Missing analysis and imputation

Outlier analysis and treatment

Encoding techniques

Binning algorithms

Custom components for specified preprocessing tasks

Model training

Hyperparameter tuning techniques

Training of supervised and unsupervised models

Ensemble methodologies

Interpretability

Analysis of variable importance using advanced interpretability techniques (SHAP, LIME, Surrogate models)

Univariate and bivariate analyses

Tracking of the entire development process

Automatic generation of model reports

Automated machine learning

Generation of user-created components

Determination and optimization of the critical path between components

Encapsulation of components in three levels of depth

Access to the Python code of each component

Features of ModelCraft™

ModelCraft™ facilitates the application of machine learning techniques through an intuitive and robust platform, deployed as a technology that is simple, flexible, and scalable

User

Ease of use:

The ModelCraft™ interface supports the visual construction of modeling workflows, providing structured access to all parameter settings.

Personalization

Users can design their own components and modelling flows, with access to the generated Python codes.

Explainability

With a single click, ModelCraft™ produces a detailed summary of the model, including its variables and the entire development process, explaining user decisions (e.g., model hyperparameters) and algorithmic choices (e.g., reasons for variable exclusion).

Evolution

ModelCraft™ is continuously updated to meet industry demands; for example, upcoming additions to the tool include the use of large language models (LLMs) and the development of advanced neural networks.

Traceability

The platform ensures complete traceability of developments, offering full access to code, along with comprehensive logging and traceability features.

Optimized code

Each component within ModelCraft™ is fine-tuned to optimize computational efficiency, utilizing libraries from cutting-edge programming environments.

Adaptability

ModelCraft™ has additional modules to solve specific problems with ad-hoc solutions such as: ModelCraft™ University, ModelCraft™ Virology, or ModelCraft™ Validation Workspace (MVW).

Deployment flexibility

ModelCraft™ can be deployed either as an on-customer-cloud tool or as software-as-a-service (SaaS) in the Cloud, according to customer needs.

Scalability and parallel computing

ModelCraft™ is scalable and takes full advantage of parallel computing capabilities, allowing the processing of large volumes of data and executions with a high computational load.

Easy deployment

The designed architecture allows ModelCraft™ to be deployed in a very short time, including an analysis of customer needs, configuration preferences and necessary customizations.

Reduced support

ModelCraft™ requires reduced IT support, and maintenance services include the deployment of new releases.

Structure of ModelCraft™

ModelCraft™ has two primary work areas: the data management panel and the modelling panel

The data management panel enables users to connect to any database and perform tasks such as extracting, merging, and loading tables using SQL. Through the component-based modelling panel, users are able to perform exploratory data analysis, prepare data for modelling, and train and interpret advanced machine learning algorithms.

Data management

Database management
Simultaneous connection to different data repositories including MySQL databases
SQL programming

Modelling

Data processing and EDA
Modelling components and advanced machine learning algorithms
Interpretability and automatic reporting components

The modeling components are organized into three levels of depth: from level 0, which covers minimal executable tasks such as outlier detection or model training, to level 2, which involves aggregating components to construct an end-to-end modeling process.

Modelcraft Structure

IT architecture

ModelCraft™ has been developed cloud-natively, using a Kubernetes-based architecture that enables rapid deployment of the component library and a visual programming frontend.

The ModelCraft™ architecture is focused on different stacked component libraries, allowing the use of the components as microservices, and includes a user, storage and execution manager.

ModelCraft™ Library

ModelCraft™ Library is a Python/PySpark library built on reusable analytical components that are tested during development. It enables the structured, standardized, and accelerated development, validation, and deployment of models through robust, reproducible, and auditable pipelines

ModelCraft™ Library evolves ModelCraft into a lightweight, reusable library with the same components, logic, and quality, but without the need for a full platform deployment, minimizing technological dependencies.

It is built on PySpark, which allows it to be used within big data platforms already available in some organizations. This enables teams to work with massive datasets and distributed computing environments without changing their existing data environment.

The library brings together ready-to-use components across all stages of the pipeline: data processing, feature engineering, training, evaluation, explainability, validation, and monitoring. Everything is consistently structured and encapsulated to avoid repetitive tasks and to accelerate both development and validation. ModelCraft™ Library can be used as an accelerator in projects, integrated directly into the client’s environment, or as part of a co-development model, creating customized corporate libraries based on our existing components.

ModelCraft™ Library completes end-to-end model lifecycle coverage, enabling the development of Machine Learning models and standardizing their creation and deployment within a single environment. It prepares data, trains models with bias control and both local and global interpretability, and deploys them to production using native formats such as Pickle and ONNX while adhering to MLOps principles.

In addition, the IRB and IFRS 9 statistical validation module evaluates distribution, predictive power, and heterogeneity, providing metrics and statistical tests segmented by relevant time periods. Connected to generative AI tools (GenMS NEXUS), it automatically generates validation reports aligned with methodological guidelines, reducing preparation time and ensuring standardized, traceable documentation.

Credentials

Management Solutions has extensive capabilities and references in the field of advanced modelling, combined with in-depth knowledge of technology and business processes

Management Solutions provides differential value in the field of advanced modeling:

It has extensive experience in the development, validation and implementation of models and algorithms applied to multiple business areas in different industries and sectors.
It has quantitative capabilities and resources with a high degree of knowledge and specialization (two thirds of its staff are STEM).
Specifically, it has teams of data scientists who carry out projects in artificial intelligence and machine learning, as well as an R&D team dedicated to the practical application of AI applied to management.
It has developed a Chair in Big Data and Analytics together with the Universidad Politécnica de Madrid, where it develops lines of research on modelling techniques and artificial intelligence. This has allowed it to develop ModelCraft™ and keep it up to date with the latest advances in modelling and computing.

Thanks to these differential values, Management Solutions implements ModelCraft™ in all its industries and in all the countries where it operates.

Financial entities

Energy & Utilities

Telecommunications

Other sectors: manufacturing, pharma, logistics and transportation, tourism, professional services, education, construction and infrastructure, government