Yavendio is an AI company building intelligent AI sales agents for WhatsApp, focused on e-commerce businesses across Latin America. With over 7,000 customers and offices expanding from Peru into Brazil, the company has raised several million dollars to fuel its growth in one of the world's most active WhatsApp markets.

As Yavendio grew, their engineering team faced a common scaling problem: they needed the power of Kubernetes but lacked the specialized expertise to manage it.

"We had a really inexperienced team in the early stages," explains Terry Cruz Melo, Co-founder and CTO at Yavendio. "Kubernetes and all its implications are a big pain for us. Great DevOps engineers are not as easily accessible in the region."

The team was managing 60-70 repositories and needed a platform that could handle complex orchestration without requiring deep Kubernetes knowledge.

They also wanted to leverage existing AWS and GCP credits from startup programs rather than paying for additional cloud resources.

Terry discovered Northflank through an AI-powered search (very fitting for an AI company).

"I was looking for something similar to Vercel, but for backend-oriented workloads. We found Northflank and I was really inclined toward it because of the good UI and Developer Experience. It was more friendly than others tools I’ve tried.

The one-to-one support we experienced was definitely a selling point. It was super easy to talk with the team directly whenever we had questions."

Since they have a lot of AWS and GCP credits from startup programs, they wanted to drain those directly without paying for additional cloud resources. Northflank's Bring Your Own Cloud model lets them do exactly that.

Today, Yavendio's entire engineering team of 15 developers relies on Northflank for their infrastructure. Using Northflank's Bring Your Own Cloud (BYOC) model, they deploy to their own AWS and Azure accounts, maximizing the value of startup credits while getting enterprise-grade orchestration.

The platform handles their complete CI/CD pipeline: when engineers push code to GitHub, Northflank automatically builds containers and deploys them to production on AWS and Azure.

"Northflank simplifies everything. Developers don't have to worry about Kubernetes configuration, they only worry about moving their service to production as fast as possible. Someone wants to deploy a new service? They go to settings, configure environment variables, do a few clicks, and it's done."

• Multi-cloud deployment: Running workloads across both AWS and Azure while draining startup credits

• Bring Your Own Cloud (BYOC): Deploying to their own AWS and Azure accounts, letting them fully utilize startup program credits

• Automated CI/CD: GitHub integration that builds and deploys without custom recipes

• Horizontal auto-scaling: Critical during high-traffic periods like the week before Christmas when traffic spikes dramatically

• Zero-downtime deployments: "Previously we'd be down for five or six minutes during updates. Now we can push updates continuously without affecting availability. That used to be really painful, we'd lose clients."

• Private networking with Tailscale: Keeping services off the public internet for security

• Open source deployments: Running tools like Metabase, Open WebUI (as a ChatGPT alternative using their Anthropic and cloud credits), and Langfuse for AI observability

Yavendio deploys a polyglot architecture on Northflank:

• Languages: Python, TypeScript, Golang (previously Rust and Elixir)
• Frameworks: FastAPI, Hono
• AI tooling: LangGraph agents, custom OLIVE framework for agent tool sharing

The team even built their own agent protocol called OLIVE (an HTTP-based alternative to MCP) that's deeply integrated with their Northflank-powered Kubernetes infrastructure.

When evaluating an enterprise AI orchestration platform, Yavendio realized they'd already built equivalent capabilities on Northflank.

"We told them, 'We have exactly the same thing on Northflank.' They were really impressed because they'd never seen something so similar built so quickly.”

The platform has enabled Yavendio to:

• Scale infrastructure to match traffic demands during peak commerce periods
• Fully leverage cloud credits through BYOC instead of paying for managed infrastructure
• Deploy continuously without service interruptions
• Empower all 15 engineers to manage deployments without DevOps expertise
• Maximize value from cloud credits across multiple providers

Yavendio is proof that a 15-person engineering team can run sophisticated infrastructure across multiple clouds without a dedicated DevOps function.

As the team expands from Peru into Brazil, their Northflank-powered infrastructure is ready to scale with them. The BYOC model means they can spin up resources in new regions without migrating platforms or renegotiating contracts.

Plus, their lean engineering team can stay focused on building rather than wrangling Kubernetes.