About this role
Forward Deployed Engineer, AI Staff — MSPbots
Most MSPs know their processes are a mess. They can't write a usable SOP, don't know which of their workflows would actually make a good AI agent, and won't self-serve their way to one.
That's where you come in.
You'll sit inside the client's operation, extract the real process from the chaos, and ship a working agent on our stack — in days, not months.
What you'll actually do: 🔹 Run discovery sessions to find the 1–3 processes worth automating (and say no to the rest) 🔹 Turn a messy, undocumented process into a real spec 🔹 Build the agent on our deepagents-based harness — tools, subagents, memory, human-in-the-loop approvals 🔹 Wire it into the client's real data, safely, with tenant isolation intact 🔹 Run shadow-review until the client trusts it, then ship the runbook and train their team
Who we're looking for: ✅ 2+ years in MSP operations (or 1+ year at MSPbots in a client-facing/delivery seat) — you can read a ticket flow or collections process and know where the pain is ✅ Code-literate — you don't need to write Python from scratch, but you need to catch when AI-generated code or SQL is wrong even though it runs ✅ Comfortable with ambiguity — you can turn "our dispatch is a mess" into a spec without being handed a perfect ticket ✅ You've automated something yourself before, even informally — a script, a workflow rule, a bot ✅ Confident on camera with a skeptical client, including when the answer is "this isn't ready to automate"
✅ Willing to work between 8 AM - 5 PM CST