Learn to build agents.
In Atlanta. Hands on keyboards.
Hands-on Claude and agent workshops for teams in Atlanta. No theory days. No fifty-slide intros to what an LLM is. The room ships a working agent, writes its handbook, and walks out knowing how to keep it alive on day two.
Five modules.
Everything that actually ships agents.
Drawn from the same playbook we use to put agents into production for clients. Operator-readable, engineer-respectable, and short on filler.
Claude Agent SDK
Hands keyboards. Build a working agent end-to-end on the Agent SDK. Tools, memory, streaming, error states. The room ships, not watches.
Agent handbooks
Write the rules the agent follows. Voice, scope, escalation paths, refusal patterns. The artifact your ops team reviews before the agent goes live.
Security & the lethal trifecta
Prompt injection. Tool-use leaks. The lethal trifecta, local data, untrusted content, outbound network, and how to split an agent so it never sits inside it.
Governance baseline
Map your agents to EU AI Act, NIST AI RMF, and ISO 42001. Audit trail, human-in-the-loop gates, vendor review. The paperwork that lets you actually ship.
Day-two operations
What happens after launch. Eval cadence, drift detection, ops manual, on-call. The unglamorous half of agent work that decides whether an agent stays in production.
From operators to engineers.
Same room.
The best workshops we run are mixed. The people who own the workflow and the people who own the code, learning to ship against each other instead of around each other.
Operators & ops leaders
Owners of the workflow being automated
You don't need to write code. You need to know what an agent should and shouldn't do, how to write its handbook, and how to supervise it once it's live. We teach that in plain English.
Product & program managers
Owners of the roadmap
You're scoping agent work for the first time. Learn to spec an agent the way you spec a feature, success metrics, scope, failure modes, so engineering can actually ship against it.
Engineers & technical leads
Owners of the build
Hands-on the Claude Agent SDK. Tool design, RAG access controls, security patterns, evals. You leave with code in your repo, not slides in your inbox.
Workshop. Cohort. Retreat.
Pick the shape. We scope the rest.
Three default shapes. Every engagement gets tuned to the team in the room and the agent you're trying to ship.
Intensive
A full day with a leadership team. Shared language, a working agent prototype, and a 90-day plan you can walk into the next exec meeting with.
Cohort
Multi-week cohort for a working team. Weekly sessions, async homework, real agents shipped against real workflows. Ends with a portfolio review.
Offsite retreat
Take the team out of the building. Mixed operator + engineer room, deep focus, working agent in your environment by the time everyone goes home.
Tell us who's in the room.
We'll scope the workshop.
Email chris@yndr.com or book a call. Rough headcount, rough mix of operators and engineers, and one sentence about the agent you're trying to ship. We come back with a format and a date.