MaxinAI IDentity

KYC / Identity Verification Solution


  • Document data extraction, validity check
  • Document type detection
  • Face recognition, comparison and matching
  • Face search
  • Other biometric checks
  • Customization

Document data extraction

  • Last Name and First Name
  • Date of birth, place of birth
  • Document ID
  • Personal ID
  • Issue date, valid date, validity check
  • MRZ* data including all data
  • Document photo area detection

Document type detection

  • Detect type:
    • ID card
    • Driver License
    • Passport
    • Other document types
  • Remove background from photo with document

Document type detection

  • Detect faces on images
  • Extract face features (feature vectors) for further processing
  • Compare two face feature vectors and predict similarity
  • Match faces on different photos
  • Store face feature vectors in a database for further use
  • Predict different face features: Hair color, style, face emotion, etc.

Face search

  • Search similar faces in face feature vector database
  • List of similar faces - top N
  • Face search in video stream
  • Person (face) tracking in video stream

Other biometric checks

  • RFID integration
  • Fake face detection (experimental)
  • Fingerprint similarity check (experimental)
  • Voice identification (experimental)
  • Document validity check (experimental)


  • Customize under your needs
    • Face features, thresholds
    • Document types
    • Biometric checks
    • REST API, RPC, Lambda, WebAssembly, Mobile SDK
    • Deployment options - Cloud, SaaS, on-premise

Cooperation stages

NDA and Agreement

Sign legal documents

Data, requirements

Basic data is provided: Type of documents, Sample images, required data-points

Data analysis, SoW

Data is analysed by our team, scope of work is created

Implementation, fine tuning

We fine-tune existing models, do changes according to requirements. Add new modules if necessary

Delivery and deployment

Deployment process according to requirements


Price depends on:

  • Modules and features
  • Document types
  • Deployment type
  • Integration requirements
  • SLA/support conditions
Email [email protected] for details

Know Your Customer


This document is a strictly confidential communication To and solely for the use of the recipient and may not be reproduced or circulated without MaxinAI Ltd.’s prior Written consent. If you are not the intended recipient, You may not disclose or use the information on this document in any way. The information is not intended as an offer or solicitation with respect to the purchase or Sale of any security.


Steps of collaboration

Procedure for retail and merchant customer onboarding

  • Step 1 - Filling Mobile/Web application;
  • Step 2 - Sending back the signed contract and details;
  • Step 3 - Customer verification and activation;

Big Data pipelines for customer data aggregation and fast analysis

Automated processing of customer data

Steps of Collaboration

NDA and First Phase Contract

Signing NDA and contract for the Phase One collaboration

Phase One Technical Tasks

Agreement on Technical Tasks and Deliverables for the Phase One

PoC for Phase One Tasks

Development of the PoCs of the tasks described in the scope of Technical Tasks Agreement

Implementation and Testing

Deployment and testing of the modules developed in the scope of the Technical Tasks Agreement

Integration in Production

Integration of the tested modules with already in production applications

Procedure for Retail and Merchant Customer Onboarding

Filling Mobile/Web Application

Sending Back Signed Contract and Personal Data

Customer Verification and Activation

Filling Mobile/Web Application - Step 1

Application is capable of extracting the information from three types of ID documents
  • Drivers License
  • International Passport
  • ID Card

Captured Data

  • First Name
  • Last Name
  • DOB
  • ID Number
  • MRZ Details
  • Majority of application fields will be filled automatically
  • Audio data is used as an additional information for verification in the future
  • Video information allows to check aliveness of the person and reduces risks of frods
  • Customer should enter the postal address and submit the application
  • Optionally GPS location information of the mobile device can be captured

Sending Back Signed Contract and Personal Data - Step 2

Request coming from Mobile/Web app

contract sent to the postal address for filling and signing it

Paper based contract is a good source of getting additional information about the future customer, also safe legal mechanism for avoiding personal data related issues;

Paper based communication is also efficient, when there is no digital state citizen verification system;

Procedure for Retail and Merchant Customer Onboarding - Step 3


Once the signed contract is received and data is validated customer gets activation SMS, which should be entered in the application

Video & Voice Verification

After entering the SMS activation code, application will be launched and request to take a short video and record voice to verify the customer according the prior data, received during the Step 1


Once customer is verified using over video and voice, account will be activated

Big Data Pipelines for Customer Data Aggregation and Fast Analysis

  • Each newly registered and activated customer will be automatically stored in two BigData “storage” systems - Message buffer e.g. Kafka - NoSQL database, e.g. HBASE
  • Deployed Machine Learning Models will be constantly retrained on the database data and provide following functionality:
    - Customer segmentation
    - Customer Churn Calculation
    - Propensity Score Matching

Automated Processing of Customer Data

With Aggregated Data it is possible to build and deploy

  • Customer behaviour models - ML regression and probabilistic models
  • Complex customer segmentation/grouping models according to large number of parameters - ML clusterization model
  • Transaction segmentation and analytics - ML classification problem
  • Campaign/discount planning - ML time series analysis problem

All these and many more models can be automatically deployed, retrained and adjusted to the live data, i.e. perform stream data analytics to achieve:

  • High Conversion Rate
  • Growth of loyal clients
  • Performing targeted and highly efficient campaigns
  • etc.

Please leave your email so that we can contact you later