Product Lead, Rippling AI – Agent Harness & Runtime

Responsibilities

• Own the architecture and product direction for Rippling’s agent harness and runtime, including planning, tool calls, result interpretation, retries, escalation, and guardrails.
• Define working memory, episodic memory, context boundaries, persistence, compaction, and intentional forgetting.
• Build the self-improvement loop through evaluations, trajectory review, and counterfactual benchmarking.
• Define improvement metrics including task completion, human-escalation rate, cost and latency per completed task, and evaluation regressions.
• Guide sandboxing, execution environments, permission tiers, connector ACLs, credential handling, and observability.
• Extend the platform into computer-use and autonomous web/research capabilities.
• Make runtime architecture build-versus-buy decisions and write technical design documents.

Requirements

• 5+ years of combined experience, including experience as a software engineer and product manager.
• Production experience building or shipping agentic systems.
• Experience shipping production systems, debugging distributed systems, and reasoning about state machines and failure modes.
• Ability to distinguish agent frameworks from production agent harnesses and runtimes.
• Strong expertise designing evaluations and benchmarks that predict production behavior.
• Working knowledge of context-window management, tool orchestration, RAG, semantic and episodic memory, OAuth grant types, sandbox isolation, reinforcement learning, fine-tuning, and non-parametric adaptation.
• Comfortable collaborating with senior engineers and independently authoring technical design documents.

Compensation

• Annual salary: $174,000–$290,000, plus benefits and equity.