Mocks & Synthetic Data
Fast and cheap, but fixtures and local samples can't tell you how the real system behaves.
Live interview + demos
How monday.com Runs Agent Evals Against Real Dependencies
A new model or a prompt tweak can quietly break what your customers rely on. monday.com runs every PR through an eval suite pointed at a staging cluster with real services and data. We'll show you how to wire the same gate into your own CI.
Featuring Dor Cohen (AI Engineering Director, monday.com) and Eyal Bukchin (CTO, MetalBear)
Where evals fall behind
What do your evals run against?
A lot of what an agentic product does is call other systems: databases, internal services, third-party APIs. One changed prompt, model, or tool can quietly break those calls. Evals catch that, but only if the dependencies they run against behave like production.
Ephemeral Environments
Real behavior, but you pay for every run in spin-up time and money.
Real Dependencies
Your eval suite, run against a staging cluster with real services and data, on every PR.
What you'll get
Learn how to run your evals against real dependencies.
You'll see:
- The design decisions behind monday.com's agent eval setup.
- What a suite catches once it talks to a populated staging cluster.
- Two live demos: a mirrord demo wiring an eval gate to a cluster from scratch, then monday.com's own framework for running agent evals.
You'll walk away with:
- The full path: eval suite → mirrord → staging services and data → merge gate in CI.
- The demo repo and a setup guide, ready to run on your own cluster.
Who should come
Who this session is for
AI & ML Engineers
PMs & Eval Owners
Platform, DevEx & Backend Engineers
Engineering Leaders
In conversation
Meet Our Speakers
Dor is AI Engineering Director at monday.com and speaks publicly about agent evaluation, including at LangChain's Interrupt in London. At monday.com, agent PRs move through the same repository, CI pipeline, and engineering standards as human-written work.
Eyal co-founded MetalBear and leads engineering on mirrord. mirrord lets developers and coding agents run local processes with access to a real Kubernetes cluster, without deploying first. Eyal will host the conversation and demo how to set up evals with mirrord.
FAQ
In case you're wondering
Is this only for Kubernetes teams?
The demo runs on a Kubernetes environment. The wider problem appears in any service-based system where local code depends on databases, queues, APIs, or services available only in a remote environment. During the webinar we'll share a link for those interested in seeing this work across different environments and setups. Or skip the wait and book a demo with us, mentioning which environment you're interested in.
Do I need mirrord installed before the session?
No. You'll be able to watch what happens live, then use the repo and setup guide we send after to run the same checks on your own cluster.
Can you eval an agent against staging without deploying it first?
Yes, and that is the point of this session. mirrord lets you run the eval locally or in CI while the agent connects to real databases, queues, and services in the cluster. The score reflects production behavior before a single line reaches it.
Do we need a separate environment to run evals against real infrastructure?
That's the problem mirrord is built to solve. You run the eval locally or in CI, and mirrord connects it to the real cluster without duplicating it. Multiple agents and engineers can run against the same environment at the same time without collisions and without the cost of a parallel staging setup.
We already run evals with LangSmith, Braintrust, etc. Why attend?
This session isn't a new eval framework, it's about what your existing suite runs against. You'll see how to point the same evals at a real database and real services, and how to wire that run into CI.
What if I can't make it?
Register anyway. We'll send the recording, slides, and demo repo to your inbox.
Event access
Run agent evals against real dependencies.
Tuesday, Sep 15 · 10am ET / 4pm CET
Free and live, plus a Q&A.
Claim your free seat