mlytics-careers

We turn content into commerce. Now turn your ideas into shipped product.

View the Project on GitHub mlytics/mlytics-careers

← Back to all positions 繁體中文版

Senior Backend Engineer, Decisive Engine

Mlytics in 30 seconds

Mlytics builds infrastructure and AI products for media and content platforms. We started with multi-CDN traffic orchestration, and the real-time decision systems behind it now handle PB-scale traffic with <50ms routing and multi-vendor failover.

Today, we are extending that foundation into AI-powered content and decision products. The engineering problems are concrete: keep real-time decisions fast and correct, move high-volume data reliably, expose capabilities through clean interfaces, and make the infrastructure economically sustainable.

More about Mlytics →


The role

The Decisive Engine is one of the most valuable systems at Mlytics. It is the backend decision layer behind our multi-CDN product: real-time routing, scoring, contextual bandit logic, attribution, and the Kafka / Redis / ClickHouse data path that keeps those decisions moving.

This is not a role where you receive a clean service and add endpoints around the edges. You will become the primary technical owner of a complex, proven system operating at PB-scale traffic — understand why it was built the way it was, keep what works, challenge what no longer does, and make it faster, safer, and more capable.

The next chapter is not maintenance. We want the engine to become accessible to the agent ecosystem: OpenAPI-first today, MCP-compatible interfaces and A2A possibilities next. Your job is to deepen the backend we already have while creating the interfaces it will need in an agentic world.


What you’ll actually do

Take ownership of the core decision engine — then move it forward.


End-to-end means production too

The person who designs a decision path should also know how it behaves under real traffic. You will share on-call responsibility for the services you own, respond when the decision pipeline degrades, restore service, and turn the incident into a durable improvement — code, monitoring, automation, or a runbook.

This is not an operations-first position, and success is not measured by how many fires you fight. Production ownership is the feedback loop that makes you a better backend engineer: it exposes the failure modes, performance cliffs, hidden coupling, and cost assumptions that architecture diagrams miss.


What we’re looking for

The non-negotiables:

What would make you exceptional:

You might not be a fit if:


Tech stack

Layer What we use
Core services Go / Python
Decision systems Routing / scoring / contextual bandits / attribution
Data path Kafka / Redis / ClickHouse / PostgreSQL
Infrastructure EKS (Kubernetes) / GCP / AWS / multi-cloud CDN architecture
Observability Prometheus / Grafana / Sentry / Loki
Agent interfaces OpenAPI / MCP-compatible services / A2A exploration
AI tooling Claude Code / Cursor / internal code generation tools

The tools are real, but the job is bigger than the stack. We need someone who can trace one routing decision across services, data, and infrastructure — and still explain why that decision matters to the product.


What success looks like


Why this matters right now

Mlytics began with multi-CDN routing. That decision engine is not legacy to be kept alive while the company moves on — it is the backend foundation the next product is being built on.

The engineer in this role will decide whether that foundation merely survives growth or becomes a deeper product capability: faster decisions, better economics, safer iteration, and a new interface for the agent era. If you want to inherit a hard system, make it unmistakably better, and open the next surface from it, this is that role.


Want to see how we actually build this?How we ship: Becoming Product Builders with Business Thinking


How to apply

Send us something that shows how you think about backend systems — a complex service you took over, a latency or concurrency problem you solved, a production incident that changed your design, or an AI-native workflow you made real. We care about what you shipped, what you measured, and what became better because you owned it.

📧 [email protected]