Founding Platform Engineer @Waniwani

Your job is to make Waniwani’s infrastructure invisible at scale.

We launched the first insurance app inside ChatGPT. Now hundreds of companies want to deploy their own AI-native storefronts through our platform. Every one of them needs its own MCP server, its own codebase, its own distribution. All running flawlessly, all interconnected, all managed by you.

Why This Role Exists

We’re past proof of concept. The product works. Now we need to run it for hundreds of customers simultaneously, each with their own MCP server, their own data pipelines, their own distribution dashboards, without the platform breaking a sweat.

This is a systems-level engineering problem. You’ll own the architecture that lets Waniwani scale from dozens of deployments to thousands. Multi-tenant infrastructure. Multi-codebase orchestration. Real-time data flowing across hundreds of interconnected services. The kind of work where one bad abstraction costs you months, and one good one buys you years.

We expect 95% of your code to be written by AI. Your value isn’t in typing. It’s in knowing what to build, how to architect it, and when the AI got it wrong. You’ll operate at the speed of someone who treats LLMs as a core part of their engineering workflow, not a novelty. The bottleneck is judgment, not keystrokes.

What You’ll Actually Do

• Own the multi-tenant platform. Architect and maintain the infrastructure that hosts hundreds of MCP codebases. Each isolated, each configurable, all observable from a single pane of glass (or MCP).

• Build the data backbone. Design the pipelines that aggregate statistics, analytics, and performance metrics across every deployed instance. Make the data real-time, reliable, and actionable.

• Ship the core app. Work hands-on in the main Waniwani application: a Next.js/TypeScript codebase with Drizzle ORM, React Query, and a modern component architecture. You’ll build features that customers and internal teams use daily.

• Design for scale. Make architectural decisions that hold at 10x. Multi-codebase deployment strategies, database partitioning, caching layers, job queues. Whatever the problem demands.

• Automate everything. CI/CD pipelines, codebase provisioning, health monitoring, rollback systems. If a human has to do it twice, you build a system to do it forever.

• Instrument relentlessly. Logging, tracing, alerting. When something breaks at 2am across 300 instances, the system should tell you exactly where and why, before anyone notices.

You’ll work directly with the CTO. The architecture is yours to shape. The stack is TypeScript end-to-end.