Transforming fashion: DoiT helps Pronti AI migrate to a serverless optimized infrastructure on Google Cloud
Meet Pronti AI
Pronti AI revolutionized its fashion AI app, achieving enhanced security, agility, and cost optimization with DoiT on Google Cloud
Pronti AI revolutionizes how people engage with fashion by harnessing the power of artificial intelligence. With its innovative mobile application, Pronti AI provides users with tailored outfit recommendations and expert shopping guidance derived from their personal closets, specific occasions, and even their current moods. By allowing users to upload pictures of their clothing items effortlessly, Pronti AI unlocks a world of fashion possibilities, empowering individuals to make informed style decisions and confidently express themselves through their unique sense of fashion. To enhance the experience, Pronti also gives suggestions on items to buy that fit the clothes the user already has.
The Challenge
Pronti AI initially developed its fashion stylist app on AWS, but struggled to scale and right-size resources. As initially set up, Pronti AI’s infrastructure was over-provisioned and lacked autoscaling. Additionally, its self-managed Kubernetes database increased risk and maintenance overhead. Together, these challenges drove unnecessary costs and operational complexity. With a small IT team and limited AWS depth, Pronti AI needed a simpler, managed approach—starting with its core cloud footprint.
After Pronti AI chose to migrate to Google Cloud and redesign its architecture, the team worked with Google to find the right partner to execute the migration. Google recommended DoiT, a Google Cloud Partner of the Year, as a full-service, multicloud partner with global presence.
The Solution
DoiT and Pronti AI started working together at the start of the migration and redesign, which enabled DoiT to pinpoint weaknesses in the previous architecture. Through its review, DoiT identified three critical areas: inflexible resource provisioning, heavy maintenance, and gaps in security. DoiT proposed a simplified Google Cloud architecture that maximized managed and serverless services, shifting operational burden from Pronti AI to the platform and DoiT.
Given rigid provisioning, DoiT first focused on scalability and elasticity. By adding Google Cloud Run, Pronti AI could absorb demand spikes without managing infrastructure. Integrating Cloud Run with Google Cloud Build streamlined CI/CD and autoscaling for APIs, ensuring responsive capacity for varying workloads. To ensure efficient database schema upgrades, DoiT supported Pronti AI in setting up Cloud Run Jobs for database schema upgrades. Isolated, ephemeral runs allowed safe, independent schema changes, minimizing impact on application performance and availability. Built-in scalability, efficiency, observability, error handling/rollbacks, CI/CD integration, and security in Cloud Run Jobs streamlined schema upgrades.
To automate build/test/CI/CD for APIs, DoiT’s Cloud Reliability Engineers (CREs) recommended integrating Cloud Build with GitHub. This integration kept APIs up to date and eliminated manual deployments. The integration enabled CI/CD, improving developer efficiency. Cloud Build allowed custom build steps and configurations. This flexibility let the team tailor pipelines to its requirements. Build triggers, environment variables, tests, and custom scripts enabled rich customization. And by deploying API services via Cloud Build integration with GitHub repositories, Pronti AI could now leverage Cloud Run’s built-in monitoring and logging capabilities. This integration provided insights into API performance, availability, and error rates. They could utilize Google Cloud Logging to capture and analyze logs for efficient debugging and troubleshooting. By leveraging GitHub’s popular version control platform and integrating it with Cloud Build, Pronti AI gained the benefits of version control for its API services, such as branching, pull requests, and code reviews. The integration streamlined collaboration and change management and ensured a well-documented history. DoiT’s strategic guidance enabled the small IT team to ship faster with higher quality through customized automation.
To finalize the migration and optimization, DoiT hardened the security architecture. To enhance security and isolation, DoiT moved the database to internal IPs and connected it via VPC Serverless Connector. This approach reduced attack surface, improved latency, simplified networking, and eased compliance, while integrating cleanly with VPC resources.
The Results
After implementing the solution guided by DoiT, Pronti AI achieved key business outcomes, including:
- Increased Security and Agility
The new architecture strengthened security, reducing breach risk. The more agile infrastructure allowed Pronti AI to deliver changes faster, improving system robustness. - Automated Testing and Deployment
The newly architected infrastructure enabled automated testing and deployments, improving developer efficiency and release reliability. - Cost Optimization
By leveraging managed and serverless services, Pronti AI right-sized compute to demand, eliminating over-provisioning and reducing operational complexity and costs.