Agent Harness Engineer - Core Runtime
Core
Build the core runtime, orchestrator, and monitoring layer that turns LLMs into reliable, autonomous software agents, managing context, tool calls, and long-running processes.
Role type
Senior IC backend/systems engineer (autonomous AI agents)
Builds
The agent harness runtime, orchestrator, and reliability layer for the SuperNinja platform
Domain
Generative AI, autonomous agents, systems engineering
Deliverable
production ML models | infrastructure
Required skills
Python, concurrency, process supervision, fault-tolerant design, LLM APIs, agentic patterns, observability/telemetry, scalable backend systems
Preferred skills
Agent framework/workflow engine experience, sandboxing/containerization, reasoning about non-determinism, rapid prototyping
Technologies
Python, Claude, Codex
Responsibilities
Design and evolve the agent run loop for context management and recovery; Build the orchestrator for long-running background agents; Implement monitoring to detect and intervene in stuck or runaway agents; Own provider abstraction across model backends; Make the harness observable with structured telemetry; Design self-healing mechanisms for graceful degradation
Seniority
Senior, hands-on IC