Metflow - from AI to Decisions
PROJECTS
GNXT
4/1/20263 min read
About the customer
Metflow is an AI-native enterprise platform designed to help organizations unify data, automate complex workflows, and generate real-time insights to drive smarter decision-making across departments. By blending advanced analytics with AI-powered automation, the platform empowers businesses to transform raw, disparate data into actionable outcomes that significantly improve operational performance and enhance cross-functional collaboration.
Challenges
Solutions
Results & Benefits
Figure 1: The architecture diagram of Metflow
To deliver a production-ready generative AI platform, Metflow designed a hybrid model architecture that combines self-hosted open-source large language models with managed foundation models on AWS. This approach enables the platform to balance performance, flexibility, and cost while supporting different AI workloads through a unified application experience.
For domain-specific AI workloads, Metflow deployed the Qwen 32B foundation model on Amazon EC2 G6e GPU instances in the Tokyo Region for both development and production environments. The NVIDIA L40S GPU-powered instances provide the compute capacity required for low-latency inference while eliminating the infrastructure limitations of running large language models on-premises. By hosting the model on dedicated GPU instances, Metflow retains full control over model configuration, deployment, and optimization to meet customer-specific performance requirements.
To extend the platform's generative AI capabilities, Metflow integrated Amazon Bedrock to access Anthropic Claude foundation models through fully managed APIs. Amazon Bedrock enables the application to leverage advanced reasoning and natural language generation without managing additional model infrastructure, allowing new AI capabilities to be introduced rapidly while benefiting from AWS-managed scalability, security, and model updates. This hybrid architecture allows the platform to route requests to either the self-hosted Qwen model or Claude on Amazon Bedrock, depending on workload characteristics and business requirements.
Don’t leave optimization opportunities on the table
Partner with GNXT and maximize your cloud investment today
About GENEXTECH SOLUTIONS
By deploying the Qwen 32B foundation model on Amazon EC2 G6e GPU instances and integrating Amazon Bedrock for managed access to Anthropic Claude models, Metflow delivered a scalable generative AI platform capable of supporting production workloads. The optimized architecture reduced AI inference processing time by 60% while maintaining stable performance and reliably handling thousands of requests per day.
The cloud-native implementation also improved data processing efficiency. Through the automated data pipeline, the platform is capable of processing thousands of terabytes of data, reducing ETL processing time by 40% while optimizing storage utilization and operational costs.
Finally, by combining AI-generated insights with near real-time business intelligence dashboards, the solution enables users to access analytics with less than 5 seconds of latency. This significantly accelerates business operations, reducing decision-making time by 50% and allowing stakeholders to make faster, data-driven decisions.
Business Challenge
As an AI-native enterprise platform processing massive volumes of organizational data, Metflow faced critical operational and technical hurdles in scaling its AI-driven architecture. Specifically, the company needed to seamlessly scale large language models—such as Qwen 32B—to analyze extensive customer data sets with high accuracy and low latency. Without a robust infrastructure to support large-scale data ingestion, storage, and high-throughput model inference, Metflow risked severe performance bottlenecks, delayed insights, and an inability to meet the real-time demands of a fast-growing user base, directly threatening its operational scalability and market competitiveness.
Furthermore, the platform encountered the challenge of effectively translating complex analytical outputs into clear, intuitive, and actionable insights for end users. As data complexity and volume grew, failing to deliver streamlined data visualization and accessible insight delivery would lead to poor user adoption, reduced operational efficiency for client organizations, and diminished overall platform value in the competitive enterprise AI market.
Architecture Diagram
Genextech Solutions is a next-generation technology consulting firm and AWS Advanced Tier Services Partner. We specialize in AI, Cloud Innovation, DevOps, and Automation — helping businesses modernize infrastructure, accelerate digital transformation, and unlock sustainable growth.
We deliver practical, secure, and scalable technology solutions tailored to your business goals. More details
GENEXTECH SOLUTIONS
Empowering businesses with innovative cloud technology solutions
© 2025. All rights reserved.
Address
17B, Street 12C, Long Binh Ward
HCMC, Vietnam
Contacts
+84 38 569 4657
innovate@genextech-solutions.com