Agentic Engineer
Farmland LP
- Location
- Larkspur, CA, United States
- Track
- AI Agent Engineer
- Salary
- $140K–$170K / yr
- Posted
- November 11, 2026
- Source
- Indeed
Job description
Tagline: Help Build the AI Operating System for Regenerative Agriculture
Farmland LP is a rapidly growing investment firm that acquires high-quality farmland, converts it to organic and regenerative practices, and manages it to generate compelling risk-adjusted returns. We have a 17-year track record, $400 million in AUM across three funds, and more than 15,000 acres, all managed by our 60-person team. Our mission is simple: build wealth by improving the land that feeds the world.
We are building an AI-native operating system where all operational, financial and investor data flows into one governed platform and every business term has a single canonical definition. That platform is a digital twin of the business: a single, trusted picture of the firm that our agents can read, reason over and act on. We're betting that AI can take on the busywork one proven step at a time, so the firm grows while our people stay focused on the decisions that matter.
The Role
We’re seeking an agentic engineer to put agents to work across our firm, building reliable systems that automate routine work and support our growth. The systems you build will support everything from farm planning and financial reporting to the answers we provide investors.
You’ll work alongside our Accounting, Finance, Investor Relations and Farm Operations teams to learn how their work gets done and identify workflows that agents can support. You succeed when they do. You’ll develop the judgment to know what’s worth building, the engineering skill to build it, and the patience to see it adopted.
You won’t be starting from scratch. Our data lives in a governed warehouse on Microsoft Fabric, modeled in dbt and described by an ontology written in LinkML. Staff access data through our own MCP server and use custom apps we’ve built for budgeting, crop planning, harvest tracking and financial reporting.
You’ll work closely with our Head of Data & AI, who will provide technical direction, hands-on mentorship, and support as you grow into ownership of these systems. You don’t need to arrive knowing our entire stack. We’re looking for strong fundamentals, curiosity, and evidence that you build things and improve them by learning from the people who use them.
What You'll Do
Map the work, then hand it to agents. Sit with Accounting, Finance, Investor Relations and Farm Operations as they close the month, build budgets and answer investors. Capture each as a workflow where every step names its owner, the check that proves it's done, and the action that completes it. Then build the agents that run its routine steps, each one supervised until its record earns it autonomy. You're done when the team runs it without you.
Grow the semantic layer. Every decision, by a person or an agent, is only as good as the number underneath it. Our MCP server resolves each business term to one governed definition, and you'll extend it with new terms, metrics and tools. Then make it learn from use: instrument it, turn every gap into backlog and every discrepancy into a test, and expand the vocabulary and ontology as the business changes.
Replace forms with actions. Every change to our data goes through a typed action: a contract for what may be written, by whom, and what record it leaves. You'll write these actions, then put self-hosted language models in front of them, so staff can work in a sentence or a screen built for the task instead of a fixed form. The model proposes, and the action decides what's valid.
Your first project: answering investors the same hour
Investors expect precise answers, both before they invest and for as long as they're with us. A fast, accurate answer is one of the ways we earn their trust. Your first project gives Investor Relations instant access to everything we've told investors and every number behind it.
In the first weeks, you'll collect the questions investors actually ask, which become the test set. You decide with us where each kind of document officially lives, who may see what, and when the tool may quote a number. Then you build it: the documents ingested and searchable through the same connector staff already use, with answers that cite their source.
We'll know it worked when an IR team member answers a real investor question from the tool, with the source cited, without asking anyone, and the numbers match the warehouse.
Qualifications
We're looking for strong fundamentals and the curiosity to learn the rest, and we'll teach you the specifics.
· 1-3 years of professional software experience, working in Git with pull requests and code review.
· You've shipped an LLM application that real people use and supported it after its launch.
· Strong Python and SQL skills.
· Daily working use of AI tools such as Claude Code.
· Evaluation discipline: you can show how you knew your system was right, and how you knew when it stopped being right.
· Exceptional communication skills with non-technical colleagues.
· Self-motivated, curious, and driven to see your work in use.
Bonus: serving open-weight models on your own hardware; MCP servers or agents that use tools; retrieval over documents; evaluation frameworks; semantic layers or ontologies (LinkML, OWL), dbt; Microsoft Fabric; financial, agricultural or geospatial domain experience.
What Success Looks Like
The ideal candidate believes technology is only as valuable as the people who use it. You're excited to help a firm that is improving American farmland run smarter, and you measure yourself by whether your colleagues stop needing you for the thing you built.
Compensation includes base salary and comprehensive benefits.
Compensation & Location
The expected annual base salary range for this role is $140,000–$170,000, depending on relevant experience, demonstrated skills, and qualifications. The role is eligible for an annual performance bonus with a target of 15% of base salary, based on individual and company performance.
We offer comprehensive benefits, incl