GKE Infrastructure Optimization

“The collaboration with bespinian was smooth, committed, and remarkably fast. Their ability to understand our challenges and move immediately into action — even when timelines were tight — made all the difference. This is exactly the type of partnership we look for to move forward quickly, in a sustainable and pragmatic way.”

Laurent Vuarraz Voisin, CEO

Customer

TALK is a Swiss technology company based in the canton of Neuchâtel that develops an AI-powered platform for improving oral communication skills. Their mobile application analyzes speeches across verbal, paraverbal, and non-verbal dimensions, providing users with objective feedback on message quality, voice, and emotional expression. Certified by Swiss Made Software, TALK serves enterprise customers, including banks, insurance companies, and large corporations, helping teams develop impactful communication skills through a combination of AI technology and expert coaching.

Background

As TALK experienced growing demand for their platform, they needed to ensure their infrastructure could scale reliably while maintaining the security and performance their enterprise customers expect. The platform runs on Cloud Platform (GCP) in Switzerland, using Kubernetes Engine (GKE), and processes sensitive video and audio data that require robust security measures and high availability.

TALK sought an external assessment of their infrastructure to identify areas for improvement and gain confidence that their setup followed cloud-native best practices. They needed a partner with deep Kubernetes expertise who could not only identify issues but also work collaboratively with their teams to implement pragmatic and sustainable improvements.

Project Goals

The engagement aimed to achieve the following objectives:

• Conduct a comprehensive assessment of the existing GKE infrastructure
• Identify gaps in security, scalability, and operational efficiency
• Provide actionable recommendations prioritized by impact
• Implement improvements collaboratively with the TALK team
• Improve infrastructure efficiency, sustainability, and long-term maintainability while strengthening reliability

bespinian’s Role

bespinian delivered a structured assessment and implementation engagement that strengthened TALK’s infrastructure posture and operational confidence.

Infrastructure Assessment Workshop

An initial workshop was conducted to thoroughly assess TALK’s GCP and GKE setup. This included reviewing cluster configuration, networking, security policies, deployment patterns, and operational procedures. The workshop provided immediate visibility into the infrastructure’s strengths and improvement areas.

Detailed Improvement Roadmap

Following the assessment, bespinian delivered a comprehensive report with a prioritized list of improvements. Each recommendation included its rationale, expected impact, and implementation approach, enabling informed and strategic decision-making.

Load Testing

Load testing was implemented to better understand platform behavior under stress and identify potential bottlenecks. This provided concrete data on system behavior as user numbers grow and helped validate subsequent optimizations.

Robustness and Scalability Improvements

Working closely with the TALK team, bespinian enhanced the resilience of applications running on GKE. This included implementing appropriate resource requests and limits, configuring horizontal pod autoscaling, improving health checks, and ensuring graceful handling of pod disruptions.

Security Hardening

The overall security posture of the infrastructure was reinforced. This covered network policies, pod security standards, secrets management, and access controls, ensuring the platform meets the stringent requirements of enterprise customers handling sensitive communication data.

Simplification and Efficiency Gains

Opportunities were identified to simplify the infrastructure and optimize resource usage. These improvements reduced unnecessary complexity, improved operational clarity, and supported a more sustainable and responsible use of cloud resources over time.

Technologies Used

• Cloud Platform: GCP - Switzerland
• Container Platform: Kubernetes Engine (GKE)
• Backend: Python
• Load Testing: Locust
• Infrastructure as Code: Terraform