Built by data scientists and subject matter experts with 100+ years of collective experience

Precision You Can Trust. Decisions You Can Explain.​

Crafted by experts with 100+ years of collective experience in predictive AI and risk model development.​

Common Challenges

encountered by Data Scientists in collaboration to develop Machine Learning (ML) 

01

CONSISTENCY

Maintaining consistency and reproducibility.

Difficult to ensure that all team members are using the same technique and modeling methodology, and ensure consistently reproducible results.

02

TECHNICALITY

Dealing with technical challenges & manual work.

Working with complex algorithms, various statistical checks and ML Ops deployment, technical challenges and manual processes that can slow down progress.

03

VERSION TRACKING

Model versioning and tracking

Managing different versions of scorecard and machine learning models developed by multiple data scientist and tracking changes requires proper versioning and documentation.

Built by data scientists and subject matter experts with 100+ years of collective experience, addressing these challenges.

Datalis is a guided machine learning tool to enable citizen Data Scientists to develop, collaborate, deploy, and monitor ML models with no-code and minimal intervention – yet configurable for advanced users

01

Provide standardized Machine Learning development process

Each step of ML modelling has been standardized based on best practice.
02

Enable faster and easier model documentation

The output and analysis are automatically generated by the system, which can be downloaded for documentation.
03

Allow team collaboration

Modeling projects are stored and developed in the centralized server or in cloud, so they can be shared and used by several data scientists or team.
04

Improve model development productivity powered by GenAI Assistant

Datalis boosts productivity by enabling easier data manipulation and interpretation of complex statistical outputs through GenAI assistance.
05

Deployment flexibility

It can be deployed on cloud, making it accessible at anytime and from anywhere. Also, with the high availability nature of cloud infrastructure, data will not be lost and server up time is guaranteed by the cloud provider.

Key Functionality Features

Data Preparation & Exploration

Ensures data sampling quality through missing value handling, exploratory analysis, transformations (normal/log/encoding), and proper data splitting for model training and model validation.

Feature Engineering & Selection

Applies statistical and machine learning techniques (IV, correlation, Stepwise, Lasso/Ridge, Chi-Square, Information Gain, etc.) to derive the most predictive features.

Model Development & Assessment

Supports various models (logistic regression, decision tree, random forest, XGBoost, regression, etc.) with evaluation using classification and regression metrics, along with validation methods (cross-validation, backtesting, in-time validation, out-of-time validation).

Model Explanation & Monitoring

Provides interpretability through variable importance, SHAP, residual and breakdown plots, and monitors performance with risk ranking, PSI, KS, AUC, Gini, Accuracy, and more.

Reject Inference & Model Enhancement

Improves prediction on rejected applications using augmentation, parceling, and weighting techniques to build fairer and more accurate models.

Automation, Deployment & Documentation

Enables automated documentation (English & Bahasa), seamless API deployment and API service monitoring to accelerate model implementation.