Lead AI Engineer, Agent Systems for Orchard IoT

The Electric Plant Co; San Francisco (CA); On-site; USD 215,000 - 235,000; Full time

Job summary

The Electric Plant Co. is building a new category of plant and tree intelligence by deploying IoT hardware and a foundation model to deliver real-time insights about plant health, growth, and stress.

We are a small, fast-moving team in San Francisco exploring biology, hardware, AI, and IoT. This senior, individual-contributor role sits between the foundation model and users, shipping the client context layer, insights, and durable, reusable workflows to growers in the field and later in reusable

Qualifications

• Built and measured production AI retrieval and context assembly with quantified usefulness.
• Shipped deterministic code and used model calls only where necessary.
• Implemented multi-tenant isolation with per-client credential scoping and data deletion.
• Designed a config or skill system usable by non-engineers.
• Operated an agent system in production with drift and cost considerations.
• Built durable, idempotent, observable scheduled or event-driven work.

Responsibilities

• Client context layer and retrieval built and measured; ensure reliability.
• The tool schema exposing the platform's data and insights, including uncertainty in payloads.
• The product's agent systems — versioned, testable artifacts and their scheduled work.
• An insight pipeline rendering to chat, email, PDF, and API.
• Context schema and card structure enabling tenant isolation and data protections.
• Actuation: safely executing client-set policies with full traceability.
• Eval suites, drift monitoring, and per-client cost accounting.

Job description

The Electric Plant Co. is building a new category of plant and tree intelligence by deploying IoT hardware and a foundation model to deliver real-time insights about plant health, growth, and stress.

We are a small, fast-moving team in San Francisco exploring biology, hardware, AI, and IoT. This senior, individual-contributor role sits between the foundation model and users, shipping the client context layer, insights, and durable, reusable workflows to growers in the field and later in reusable