Q2 R&D Newsletter

A centralized solution that collects and unifies client data from multiple sources, providing businesses with valuable insights for personalized marketing and improved client experiences.

• Collect and unify client data from multiple sources
• Dynamic & Predictive Segmentation
• Personalized experiences & Recommendation
• Omni-channel Marketing Automation

• AWS S3, AWS DynamoDB, MySQL
• AWS Glue, AWS Emr, AWS Lambda, AWS SDK, Kinesis Data firehose
• Java Spring Boot, Python, ReactJS, SQL, Pyspark, JOOQ, Flyway

• 360-Degree client View
• Personalized client Experiences
• Improved Marketing Efficiency
• Enhanced client Engagement and Retention

The AI box, which can be an embedded computer, mini computer, industrial PC, or workstation, is capable of processing real-time camera feeds. It utilizes active AI processing, suitable for situations need continuous frame of the camera

The AI system can serve to cater to various functions, including: People counting in area, People counting in/out, Heatmap analysis, People counting in area, Vehicle counting in/out, Parking lot, License recognition (parking lot), Animal detection, Children detection, OCR, Fail Detection, Object Counting, Object Tracking

• Object Detection
• Object Tracking
• Multiple Camera Processing
• Computer Vision
• Image Processing
• OCR

Our AI technology can assist clients in solving various vision problems across multiple fields. These include security applications such as home, building, and factory surveillance, social applications like crowd warning, and safety applications like monitoring restricted areas in factories and industrial park

The cloud-based AI platform is capable of utilizing multiple hardware resources to process AI requests for user-uploaded images. The cost incurred from each AI request benefits the hardware owner, which can be a TMA or any other distributor

• OCR Medical Devices
• OCR ID Card, Passport
• OCR In-body, Receipt, Prescription
• Face Recognition
• Age Gender Detection
• Face Selfie and Face on Document Comparison

• Object Detection
• OCR
• Image Processing
• Computer Vision
• Template Matching
• Deep Learning

The cost associated with each AI request benefits the owner of the hardware, which could be a TMA or any other distributor. This supports multiple sectors such as clinics, hospitals, banking, building, and manufacturing

• A lightweight RPA framework
• Easy to customize
• Available AI library

• Centralized governance & monitoring
• Cross application and legacy system integration
• Intelligent automation powered by TMA AI/OCR/ML ecosystems.
• Task scheduling and prioritization
• Reporting and analytics

• Python, Django, Postgres, RabbitMQ, ReactJs, Docker
• PyQt
• RPA, Selenium, Robot Framework
• AI/ML/OCR

Available solutions for various industries

Extract the information from various types of documents

• Support structured and unstructured, printed and handwritten docs
• Reduce manual data entry and automate business
• Utilize both offline and cloud computing services to get the best result

• NLP
• Machine Learning
• Deep Learning
• Word Embedding
• Feature Extraction
• Linguistic Features
• Image Processing
• OCR

• Time-saving procedure
• High correctness
• Automatic filtering process
• Labor effectiveness
• Accuracy data extraction
• Variety of document formats

Uses AI algorithms to detect, extract text document in a \"free template mode\" and \"complex background\" from all types of documents including \"unstructured ones\"

• AI software recognizes and processes multiple languages.
• Workflow extracts relevant information quickly and accurately.
• Train templates to extract specific information.
• Document Classification automatically categorizes your documents.

• Pytorch, Tensorflow
• Opencv, Numpy
• PostgreSQL, AWS
• ReactJs
• RPA
• FastAPI, Minio

• Save Storage space: Text files are smaller than images
• Text and data extracted can be saved for further use
• Fast and Accurate Data Entry, Save Time and Cost
• Better Accessibility and Search Ability

• An end-to-end solution to predict and prevent machinery malfunctions & failure events. Our AI-backed technology can identify common root causes in advance.
• Supporting data types: Sound, Vibration and Temperature

• Anomaly detection
• Early warning failure events/downtimes
• Monitor Machine/System Health

• Python, Long-Short-Term Memory, CNN, TensorFlow, Keras, Scikit-learn
• Azure Cloud, IoT Hub, Data Lake

• Improve assets maintenance in Real-time
• Minimize unplanned downtime & optimize maintenance costs.
• Common use in industrial manufacturers & other Assest Monitoring systems

Investors around the globe can buy, sell, and swap into the market through fully-compliant, fractional, tokenized ownership. Powered by blockchain.

• Audio/Video Services

• NestJS
• ReactJS
• Flutter
• MySQL
• Solidity
• Hardhat
• Metamask
• Coinbase

• Transparency
• Users control their own data
• Data is decentralized, immutable and linked together using cryptography

A flexible Audio/Video Service designed specifically for both mobile and web applications, offering the convenience of multi-tenancy and a lightweight framework.

• Provide SDKs for mobiles and browsers
• In-call control with custom APIs
• Fast call establishment
• Good Video Quality
• API-key authentication

• WebRTC, SIP, Kamailio, STUN/TURN, Websocket
• Nodejs + ReactJS + MongoDB

• Use as an alternative to costly services like Stringee, Twilio
• Enabling video/audio capability to application with ease and quick

MEC is a network solution that provides services and computing functions required by users on edge nodes, instead of 5G core network node. It includes:

• MEC Platform (MEP)
• MEC Platform Manager (MECM)
• MEC App Orchestrator (MEAO)
• 5G Applications

• Build image of 5G Applications
• Allow deployment of 5G Applications in MEC at edge node
• Support lifecycle management of 5G Applications
• Support traffic steering in conjunction with local 5G UPF

• Edge Computing, Local Routing, Traffic Steering
• UPF (re)selection and traffic routing
• DNN, DNS, NAT
• Session and Service Continuity (SSC) modes
• Local Area Data Network (LADN)
• Open-stack, K8S, NFV

• Reduce infrastructure cost of at 5G core network node
• Allow low latency within network
• Reduce bandwidth requirement
• Improve performance of applications
• Offer new network services & new revenue opportunity

• Enable uninterrupted
• Wi-Fi connectivity for Automated Guided Vehicles (AGVs) operating in wide areas.

• Automatic scanning, monitoring, and signal analysis to connect to the best Wi-Fi network.
• Seamless message routing between interfaces during the handover with minimal quality impact.

• Wi-Fi 6 (802.11 ax), Wi-Fi HaLow (802.11 ah)
• OS: Embedded Linux, Ubuntu
• Linux Kernel/Drivers: Linux wireless driver, Linux mac80211 module
• Languages: C/C++; Python
• Protocols: TCP, UDP, SCTP, IP

• 40% higher data rate.
• 15% lower latency.
• Less than 1% connection loss, 99% guaranteed throughput for mobility handover.
• Unlimited device scalability.

• ODB2 Diagnostics Module is a system used in modern vehicles to monitor and diagnose problems.
• It collects data from in-vehicle system to report any abnormalities or malfunctions

• Read data from ECUs, sensors and electrics in-vehicle from CAN, LIN and GMLAN
• Display in-vehicle status or leading errors in applications
• Send notifications when problems occur
• Identify vehicle location

• Tools: Eclipse, VS Code, Gcc for arm, KiCad, ESP8266 NONOS SDK, AWS
• Database: PostgreSQL, OpenDBC file Encodes
• Language: Java, ReactJS, Cordova, MapBoxGL, JS/HTML/CSS, C
• Hardware: STM32F413, ELM327, ESP8266
• Basic protocol: CAN, LIN, GMLAN, SPI, Websocket

It helps vehicle owners, mechanics, and emissions inspector to identify and address problems early on, leading to improved vehicle performance, reduced emissions, and better overall vehicle health