Staff AI Infrastructure Engineer: Own the Intelligence Layer

Cassi Home; United States; On-site; USD 180,000 - 260,000; Full time

Job summary

Cassi Home seeks a Staff- or Principal-level AI Infrastructure Engineer to own the intelligence layer end-to-end—from the agent runtime to the multi-provider inference substrate and surfaces that touch users. This role drives architectural decisions, audits existing choices, and leads migrations to scale the platform with reliable, observable systems.

You will own graph topology, tool-call correctness, streaming, memory, and per-property memory across interactions, while ensuring secure,

Qualifications

• 8+ years of professional backend engineering.
• Roughly 3–5 years working directly with LLMs.
• Production experience, not demos.
• Agent and tool-calling depth.
• Inference fundamentals and streaming.
• Retrieval judgment for embeddings and memory.
• Evaluation discipline and offline/online evals.
• Security instinct for AI surfaces.
• Judgment about other people's architecture.
• Backend depth with DDD or event-driven design.

Responsibilities

• Architectural Audit & Scale: pressure-test and migrate existing design choices.
• Agent Runtime & Orchestration: manage graph topology, tool calls, state, and failures.
• Inference Infrastructure & Model Routing: provider-agnostic substrate and routing.
• Realtime Voice: manage bidirectional streaming, latency, and billing units.
• Retrieval, Context & Memory: embedding strategy and per-property memory.
• Ambient Intelligence & Evaluators: event-driven signals and evaluators.
• Evaluation, Observability & Unit Economics: tracing, cost attribution, QA gates.
• Productization — In-Product and Internal: ship surfaces behind feature gates.

Skills

TypeScript mastery

Backend depth

Database fluency

AWS

Security & reliability

Education

Bachelor's degree in Computer Science or related field

Tools

AWS

DynamoDB

PostgreSQL

LangGraph-like frameworks

Job description

Cassi Home seeks a Staff- or Principal-level AI Infrastructure Engineer to own the intelligence layer end-to-end—from the agent runtime to the multi-provider inference substrate and surfaces that touch users. This role drives architectural decisions, audits existing choices, and leads migrations to scale the platform with reliable, observable systems.

You will own graph topology, tool-call correctness, streaming, memory, and per-property memory across interactions, while ensuring secure,