Everyone’s building with AI. Most are shipping unused feature add-ons, slapped onto their products with pretty animations.
We’re not doing that. We’re building an intelligent platform that handles orchestration, evals, monitoring, and internal + customer-facing agents. We want systems that scale, tools that abstract away AI weirdness, and a dev experience that’s delightful.
You’ll be part of a small, scrappy, high-context team working across the entire LLM stack — from prompt design to infra to monitoring to customer-facing UX. We move fast, own our work, and prioritize quality + learning over throwing models at the wall.
It’s ambitious, but we think this team has the potential to build something that becomes bigger than Amplitude’s core business in a short period of time. And we have the support of the company to do that.
What will you do?
- Own core backend systems and infra for our agent orchestration layer
- Help lay the foundation of our internal AI platform (orchestration, evals, monitoring, prompting, guardrails)
- Build out individual tools and agents and workflows that interface with the rest of the Amplitude platform
- Work closely with product engineers to connect backend to UI in a seamless dev loop
- Define and implement best practices for LLMOps (tooling, evals, latency mgmt, tracing)
- Ship fast and safely — we favor momentum and reliability over bureaucracy
- Be an early voice in the architectural and product direction of this team
You might be a fit if:
- You have 5+ years of backend or platform engineering experience, ideally in high-scale, high ambiguity, or fast-moving environments
- You’ve shipped AI/LLM systems (RAG, eval pipelines, agents, etc) or are hungry to learn
- You love building internal tools and abstractions that make other devs faster and safer
- You understand the weirdness of LLMs — non-determinism, latency, token costs — and how to design around them
- You write clear, maintainable code and care deeply about testing and observability
- You like owning ambiguous problems end-to-end, not just closing JIRA tickets
Bonus if:
- You’ve worked in data or infra teams before
- You’ve built agent frameworks or LLM-based systems
- You’ve been part of a small startup or skunkworks team and thrived in uncertainty
- You’ve been a founder or early-stage at a startup
Why us?
- Scale — We’re a $300M+ ARR business and are working closely with our customers to co-develop this next product.
- Data — the existing analytics infra we’ve built puts us in a unique position to build useful AI products
- Org commitment — We are all in on this and devoting lots of company $ to it. This isn’t a sideshow research project.
- Tight team — We’re keeping the team intentionally small and insulating it from org bloat — no roadmap review, etc.
- Work closely with the Amplitude co-founders
We’re betting the next chapter of Amplitude is AI-native — and this team is leading that charge
