Containerized CI/CD Infrastructure for GoCulture’s Employee Insights Platform
Modernizing an AI-driven engagement platform with automated deployments and faster insight generation.
GoCulture delivers a science-backed employee engagement platform for actionable cultural insights, but insight generation and reporting depended on manual effort and inconsistent releases. The objective was to modernize the platform using containerized services with ECS (EC2 launch type), automate deployments with CI/CD, and enable AI-powered analysis and reporting using Bedrock and Nova Pro while preserving strict IAM and data security controls.
Challenges
- Delayed insight generation: Manual structuring and analysis caused insights to take several days.
- Lack of real-time monitoring: No automated sentiment or anomaly detection, leading to late identification of disengagement risks.
- Rigid reporting workflow: High manual effort to generate tailored reports across organizational units.
- Limited CI/CD automation: Manual, inconsistent deployments increased operational overhead and downtime risk.
- EC2 compatibility constraints: Existing dependencies required ECS (EC2 launch type) rather than a full re-platform.
The Solution
- Deployed a containerized architecture on Amazon ECS (EC2 launch type) to modernize backend and AI workflows while preserving EC2-based dependencies.
- Implemented Amazon RDS (PostgreSQL) as the central analytics database for structured and enriched employee response data.
- Integrated Amazon Bedrock for managed model execution and Amazon Nova Pro for natural language understanding, sentiment analysis, and restructuring admin questions into SQL.
- Stored raw survey inputs, transformed datasets, and report artifacts in Amazon S3.
- Delivered frontend and static assets globally via Amazon CloudFront for low-latency access.
- Automated build, test, and deployment cycles using AWS CodePipeline for consistent releases and rollback readiness.
- Enforced least-privilege access using AWS IAM across ECS tasks, CodePipeline stages, Bedrock interactions, and database access.
- Centralized logs and operational metrics in Amazon CloudWatch for ECS services, CodePipeline stages, and AI workflow performance.
The Impact
With KnackForge Cloud Services in place, the customer experienced:
- Faster release cycles: Backend rollouts reduced from several hours to under 20 minutes using AWS CodePipeline.
- Reduced manual reporting effort: Manual reporting workload dropped by 70% due to AI-generated insights and scheduled summaries.
- Faster time-to-insight: Ingestion-to-insight turnaround improved by 65%, and insight generation reduced from 3 to 4 days to under 48 hours.
- Improved risk visibility: Early detection of at-risk behavior improved by 70%, enabling more proactive interventions.
- Higher engagement: Employee survey participation increased by 30% within the first month of relaunch.
- Lower admin overhead: Overall administrative workload reduced by 35%, freeing HR staff for strategic engagement.
- Lower operating costs: A 12-month projection estimated a 20 to 25% reduction in total platform operation and maintenance costs.
Technologies Used:
- Amazon ECS (EC2 launch type)
- Amazon EC2
- Amazon RDS for PostgreSQL
- Amazon S3
- Amazon CloudFront
- AWS CodePipeline
- AWS IAM
- Amazon CloudWatch
- Amazon Bedrock
- Amazon Nova Pro
- Application Load Balancer (ALB)