About DirectHedge

DirectHedge is building the operating system for physical lumber trading. Our platform combines an order management system (OMS), an anonymous physical lumber exchange and an AI layer — DirectIntelligence (DI) — that automates operational workflows for traders and buyers.We're a small, fast-moving team in Vancouver, and what we build goes into production and is used by real users.

The Role

We're looking for an AI Engineer to join us on an initial 6-month contract, with the potential to extend.This isn't a sandbox AI project. DirectHedge already has a live MCP (Model Context Protocol) server with 30+ tools, a working OMS platform and multiple AI agents in development.You'll build on that foundation — developing MCP tools, refining agent behaviour, writing and testing agent instructions, building evaluations and connecting AI systems with production APIs.We're looking for someone who has actually built with LLM APIs and is comfortable writing code, rather than someone whose AI experience is limited to using AI tools.

What You'll Work OnMCP & Tool Development

- Build and improve TypeScript-based MCP tools
- Write tests and maintain accurate schemas and documentation
- Identify API gaps and work with the backend team to resolve them

AI Agents

- Refine agents supporting invoicing, collections, cash flow, inventory and other trading workflows
- Improve system prompts, structured outputs and tool-use behaviour
- Build evaluation workflows and test agent performance
- Help design new agents as the platform evolves

Skills & Instructions

- Write and iterate on markdown-based agent skill/instruction files
- Treat agent instructions like code: structured, tested and continuously improved

Backend Development

- Depending on experience and performance,



contribute to backend APIs supporting MCP tools and agents

What We're Looking For

- Strong TypeScript/JavaScript skills
- Practical experience building with LLM APIs such as OpenAI, Anthropic or similar
- Understanding of tool/function calling, structured outputs, prompt design and agent workflows
- Comfortable working with REST APIs, JSON and technical documentation
- Able to write clean, structured and maintainable code
- Strong written communication and attention to detail
- Able to independently investigate problems and work through solutions
- Comfortable working in a small, rapid-moving engineering environment
- Able to work in person in Vancouver, BC

Strong Signals — Not Required

- Experience with Model Context Protocol (MCP)
- Experience with LangChain, LangGraph or similar agent frameworks
- Experience building AI agents that interact with external tools or APIs
- Python experience
- Backend/API development experience
- Automated testing or LLM evaluation experience
- Git-based development workflows
- Familiarity with Linear
- Exposure to logistics, supply chain, commodities or trading software

Contract Details

- Initial term: 6 months
- Potential to extend based on performance and business needs
- Location: In-office, Vancouver, BC
- Start: As soon as possible

Why DirectHedge

- Work on AI systems already being used in a real commercial environment
- Build production MCP tools and AI agents rather than proof-of-concept projects
- Work directly with the people designing and building the platform
- Gain exposure across AI agents, APIs, backend systems and product decisions
- Take meaningful ownership in a small team where your work can have an immediate impact

Pay: $25.00 per hour Expected hours: No less than 40.0 per week

Work Location: In person

📌 AI Engineer – Agents & MCP (Vancouver)
🏢 Vantage Commodities
📍 Vancouver