ABOUT BCMR
BCMR is a Toronto-based custom construction company entering its 11th year. We specialize in high-end residential builds and renovations, with growing commercial work. We operate multiple projects at any given time, run by a small, disciplined team backed by a strong network of subcontractors across the GTA and Ontario.
We run entirely on Microsoft 365 with Copilot, and we’re building an internal ecosystem of AI agents on top of that stack — supporting estimating, internal communications, field-data reconciliation, and personal-assistant workflows. The architecture is designed and the first pilots are authored. We’re looking for someone to take what’s been planned and ship it.
THE ROLE
We’re hiring a Junior AI Engineer to build, deploy, and refine internal AI agents across our Microsoft Copilot ecosystem — Copilot Studio, Power Automate, and the wider Microsoft 365 stack. Strategy, prompt design, and feedback come from the owner — your job is execution: turning specifications, system prompts, and architectural decisions into working agents the BCMR team uses every day.
You don’t need years of professional experience, but you do need to arrive able to build. We’ll provide the architecture and direction; you bring demonstrable hands-on skill with agent platforms — whether that comes from professional work, internships, school, or personal projects. The architecture and the standards are ours to set; the build is yours to own. We work at the frontier of Microsoft’s agent platform, so an appetite to learn fast as the stack evolves matters more than a long résumé.
WHAT YOU'LL BUILD
You’ll own the build-out of a growing suite of internal agents across the Microsoft Copilot ecosystem. Specific projects will evolve, but the work falls into a few recurring categories — the examples below are illustrative, not exhaustive:
- Communications and document agents — agents that generate, summarize, or route internal communications (newsletters, status updates, project notifications), on scheduled or event-based triggers, posting to the right destination.
- Domain estimating agents — a family of agents that apply one shared architecture across cost categories, turning project data into structured outputs (e.g., Excel) and improving iteratively against real results.
- Personal-assistant workflows — voice-first or mobile capture that routes inputs to the correct artifact (tasks, email drafts, project notes).
- Data-reconciliation agents — agents that ingest operational data daily, match it against source records, and propose actions for human review.
- Copilot ecosystem extensions — as Microsoft’s agent platform matures (Copilot Cowork, Agent 365), author custom skills and plugins that bring BCMR’s workflows and knowledge into the tools the team already works in.
For each agent, you’ll handle the full path: build it, wire it into the surrounding Microsoft 365 workflow, deploy it, and refine it against live use.
WHAT WE'RE LOOKING FOR
Required
- Computer Science degree (completed or in final year) or equivalent demonstrable experience.
- Hands-on experience building agents in Microsoft Copilot Studio — not just chatbot configuration, but tool-calling agents with Power Automate flows wired in. Project, portfolio, or coursework builds count.
- Working knowledge of Power Automate — flow design, connector usage (Planner, Outlook, SharePoint, Teams), error handling, and scheduled triggers.
- Solid grasp of agentic workflow design — system-prompt construction, tool use / function calling, and how to structure a reliable multi-step agent.
- Comfort across the Microsoft 365 ecosystem — SharePoint, Teams, Outlook, OneDrive, Planner, and Excel (including Office Scripts).
- Strong written communication — you’ll document what you build so non-technical teammates can maintain it.
- Self-directed work style — remote, with minimal day-to-day supervision. We measure progress by shipped agents, not hours logged.
Nice to have
- Direct LLM API experience — Anthropic Claude, Google Gemini, OpenAI, or comparable. Copilot is multi-model under the hood, so this carries over rather than competing with the Microsoft stack.
- Exposure to construction, professional services, or another deeply domain-specific business context.
- Experience with Microsoft Graph API, Azure OpenAI, or related Microsoft AI infrastructure.
- Awareness of Microsoft’s emerging agent layer — Copilot Cowork, Agent 365, or the Frontier program — and how Copilot Studio agents fit alongside it.
- Familiarity with agentic development tooling such as Claude Code, Claude Cowork, or similar.
- Comfort translating non-technical stakeholders’ domain expertise into technical specifications.
WHAT WE'RE NOT LOOKING FOR
- A pure ML / data-science background with no agent-engineering experience.
- A generalist software developer with no exposure to LLM-powered tools or agent platforms.
- Someone looking for a rigid 9-to-5 — this role is outcome-driven, not hour-driven.
WHY THIS ROLE
For the right person, this is a chance to build a meaningful agent stack from the ground up inside a real operating business. You’re not building demos — the agents you ship get used by the team every week, and the wins are visible: time saved, fewer mistakes, faster output. You’ll work directly with the owner, who sets the architecture and prompt design, so you’re learning agent design from someone doing it daily rather than figuring it out alone. The timing matters, too: Microsoft’s agent ecosystem — Copilot Studio, Cowork, Agent 365 — is expanding fast, so you’ll build deep, early expertise in a stack that every Microsoft-based business will soon depend on. There’s room to shape what this becomes, and BCMR has long-term interest in productizing parts of this work. The right person grows with the company.
HOW TO APPLY
Send the following to brett@bcmrgroup.com with the subject line “Junior AI Engineer Application”:
- A short note (3–5 sentences) on why this role interests you and what you’ve built that’s relevant. Please name one agent or automation you’ve built and one thing you’d redo if you started it over.
- A link to your GitHub, portfolio, or any agents/projects you can show.
- Your resume.
- Your preference — hourly contract or salaried — and any constraints on availability.
We review applications on a rolling basis and respond to all candidates whose backgrounds fit the role. Strong candidates may be asked to walk us through an agent they’ve built as part of the interview process.
