3 weeks, 5 live modules.
Two evenings a week (2.5 hrs each) plus two office-hour blocks — 12.5 hrs of live instruction and ~8–10 hrs of labs and project work.
Master Claude Code and the agentic toolkit, wire AI into your tools with MCP and custom skills, deploy your own agents, and build a knowledge vault that runs your work. You ship something real every week.
Claude Code · MCP & Skills · Agents · Knowledge Vault · Evals & Economics — 12.5 hrs live
Anyone can demo an agent. We teach you to prove it works — and prove what it costs.
Each module is ~75 minutes of live instruction, a hands-on lab, and homework that builds toward your capstone: your own agent and knowledge vault, with proof they work and a number on what they cost. A non-coder track keeps every module hands-on without writing code.
Get the tools everyone is talking about working for you — and learn when to use which. Specs, plans, and subagents; skill workflows ("superpowers" like brainstorming, planning, and systematic writing) that make AI a real partner on non-coding work too.
you ship: every tool installed + your first real task done with Claude Code
What MCP is, what skills are, and when each is the right tool. Wire AI into Slack, Drive, GitHub, and your calendar; stand up a simple MCP server; then author your own reusable skill for a workflow you repeat every week.
you ship: a working MCP connection into your own stack + a custom skill you wrote
Agent loops, bounded autonomy, subagent orchestration, and scheduled agents that work while you sleep — taught from the fleet of research, lead-signal, and marketing agents we run every day at UnitSense, with guardrails against runaway loops.
you ship: your own autonomous agent, deployed and running on a schedule
The system that makes everything else compound: context files, memory, file structures agents can navigate, daily logs and handoffs. Taught on the real vault UnitSense runs the company on — and you build yours live alongside it.
you ship: your own knowledge vault, structured and driving a real workflow
Does it work, and what does it cost? Lightweight evaluations and LLM-as-judge basics, then token economics — cost per task, per outcome, per agent. Capstone presentations close the course: your agent and vault, with an eval check and a cost number.
you ship: your capstone — and earn the certification
Live, cohort-based, and assessed — a real course with defined learning outcomes and shipped deliverables, not a playlist of videos.
Two evenings a week (2.5 hrs each) plus two office-hour blocks — 12.5 hrs of live instruction and ~8–10 hrs of labs and project work.
Engineers, solutions engineers, PMs, ops, managers. No coding background required — the non-coder track keeps every module hands-on.
By the end: the full toolkit configured, a custom skill, a deployed agent, and a knowledge vault running a real workflow — not notes, infrastructure.
Graded homework and a rubric-scored capstone earn the certification — the completion record education-reimbursement programs require.
Required tools: Anthropic + OpenAI API access (~$50 in usage credits) for the duration of the course.
UnitSense builds the AI unit-economics platform — cost attribution, governance, and agent observability over live LLM traffic. The agents, skills, and knowledge vault you'll learn from are the ones running our company right now.
8+ years in enterprise infrastructure and AI systems. Formerly Director of AI Infrastructure at CTDS, where he built and operated a petabyte-scale, zero-trust Kubernetes GPU cluster for production AI workloads. MS Computer Science, University of Chicago; MBA, Chicago Booth.
10+ years in platform and infrastructure engineering. Architect of UnitSense's multi-tenant AI gateway — routing, caching, budget enforcement, and agent observability over live production LLM traffic. Previously platform engineering at CTDS, PLUMgrid, and Motorola.
The curriculum is reviewed by named external AI practitioners and refined through a rated beta cohort before each run. Reviewer list and ratings published on the course page.
Tell us a bit about you and we'll send the full syllabus, cohort dates, pricing, and the reimbursement kit as soon as enrollment opens.
No commitment — joining the list just reserves your place in line.