Staff AI Agent Runtime Engineer — Equity

Leoforce; New York (NY); On-site; USD 185,000 - 285,000; Full time

Benefits offered by this job

Equity / ownership

Direct access to founders/CTO

SoHo NYC office

Job summary

Leoforce is seeking a Staff/Senior Software Engineer specializing in Agent Engineering in New York City. The role focuses on advancing the agent runtime powering research, analysis, and action capabilities, with a emphasis on building durable, auditable multi-agent systems.

You will drive production-grade agent orchestration, memory, and model-runtime engineering across providers, ensuring secure, scalable, and observable operation in a fast-moving startup environment.

Qualifications

• Experience shipping production LLM or agent systems beyond prototypes.
• Ability to diagnose nondeterministic failures and separate model judgment from deterministic software.
• Experience with long-running, multi-agent orchestration.

Responsibilities

• Own and advance the platform's agent runtime and foundational systems.
• Build infrastructure for long-running agent work with visibility and steerability.
• Design and build the tool and action framework used by agents.
• Connect agents to retrieval and knowledge systems; ensure grounding in evidence.
• Improve persistent memory and context-management for complex workflows.
• Strengthen evaluation, tracing, and release processes for agent behavior.
• Improve quality and reliability while managing latency, cost, security, and risk.

Skills

Stateful multi-agent workflows

Tooling & contracts

Agent memory

Model routing

Observability & security

Job description

Leoforce is seeking a Staff/Senior Software Engineer specializing in Agent Engineering in New York City. The role focuses on advancing the agent runtime powering research, analysis, and action capabilities, with a emphasis on building durable, auditable multi-agent systems.

You will drive production-grade agent orchestration, memory, and model-runtime engineering across providers, ensuring secure, scalable, and observable operation in a fast-moving startup environment.