Deploy AI Agents to Production
Paramanantham Harrison
Founder @ KaaS.team, Engineering leader, Agentic AI educator
Take your agentic AI system to production in four hours.
Your agents work on your laptop. Production asks harder questions. Who is each agent acting as, and what may it touch? What stops a prompt injection hidden in a customer email? What happens when a run dies halfway through writing to the ledger? How do you know the work was really done, and what did each outcome cost? How do you know a model swap did not make it worse?
Most agent projects stall here, between the demo and the security review. In this workshop you take AI mission control through each of those questions, live: a support agent with a refund limit, and a finance agent that runs overnight.
You leave with the system deployed for a second tenant and a checklist to put your own agents through before they ship.
What you’ll learn
A one-day working session to take an agentic AI system to production: identity, guardrails, durable work, a gateway, traces and evals.
Give each agent its own identity and check every action against the user's permissions.
Block prompt injection and keep secrets out of the model.
Resume a run killed mid-way without writing anything twice.
Count an outcome only when the system of record agrees, and trace its cost.
Build evals from real traces and fail a deploy that makes an agent worse.
Leave with a production checklist for your own agents.
Workshop agenda
Identity for Every Agent
Sign each agent in with its own credentials and act on behalf of the user with Keycloak.
Permissions and Policy
Check every action with OpenFGA and write approval rules as code.
Guardrails
Block a prompt injection hidden in a customer email, and keep secrets and personal data out of the model.
Durable Work
Kill a run mid-way and resume it without writing the same ledger row twice.
The LLM Gateway
Route each step to the cheapest model that passes, cache, and budget every call per tenant.
Traces and Cost per Outcome
Trace every run in Langfuse and read the cost of each outcome, failed runs included.
Evals in CI
Build evals from real traces and fail the deploy when a model swap makes an agent worse.
Ship It
Deploy the system and onboard a second tenant live.
Learn directly from Paramanantham
Paramanantham Harrison
Founder @ KaaS.team, Engineering leader, Agentic AI educator
Who this workshop is for
Software engineers taking an agentic system from demo to production (you learn identity, guardrails, durable work and the deploy path)
Platform engineers and tech leads who will run agents for many teams (you learn gateways, budgets, traces and evals)
Engineering leaders and PMs who approve agents for production (you learn the checklist a release has to pass)
Prerequisites
You have built an agent
You have built an agent with tools, in any framework.
You know Docker and Git
You can run containers and clone a repository.
You have an LLM API key
Any major provider works.
What's included
Live sessions
Learn directly from Paramanantham Harrison in a real-time, interactive format.
Live workshop access
Four hours with Param, taking an agentic system to production.
Starter repo
The AI mission control system used in the session, to run on your own machine.
Recordings you keep
The full session, recorded.
Certificate of completion
Share what you shipped with your team or on LinkedIn.
Maven Guarantee
Your purchase is backed by the Maven Guarantee.
Frequently asked questions
Maven for Teams
Reimbursement
Get your company to pay
Everything L&D needs: email template, receipts, and certificate of completion.
Get reimbursedTeam discount
Learn with your teammates
Save 20%+ when 2 or more teammates enroll in the same cohort.
Save 20%+ with a teamPrivate cohort
Run a cohort for your org
A dedicated cohort with a custom schedule and curriculum, tailored to your team.
Book a private cohort$497
USD
10am–2pm EST