Team workshops

Your AI team ships fast. The projects still don't land.

A remote, hands-on workshop where your team runs its own AI projects through a framing process and walks out with a structured approach for their projects.

Coding agents made your engineers faster than they have ever been. A week of implementation is now an afternoon. But that speed doesn't stop you from running fast in the wrong direction. The bottleneck is not your model or your engineers. It is the framing work before anyone writes code: communicating with the business and identifying the real goal of a project. On most teams nobody owns that step.

This workshop gives your team a shared process for that step. It pairs a short overview of the method with a working session where we walk through your real use cases and rebuild the approach from first principles.

What's different from a course

It is just your team, run privately. We focus on the problems your team actually cares about, and people can put confidential details on the table without holding back. Working through your own use cases lands harder than any textbook example from Uber or Apple, because the context and the stakes are real.

What your team walks out with

• A repeatable way to frame business problems into AI solutions your team can execute.
• A diagnosis of a stalled project: is it a model problem you keep tuning, or a framing problem no amount of tuning will fix?
• A method for presenting project decisions to stakeholders, backed by evidence rather than opinion.
• A shared language, so the next argument about scope is a five-minute conversation instead of a three-week exercise.

The method

The workshop runs the same lifecycle the team reuses on every project after I leave.

• Frame with the GOATS Loop: Goal, Operating Assumptions, Alternatives, Trade-offs, Signals. Map the whole problem before committing to an approach.
• Diagnose when something stalls: tell a model failure from a framing failure in minutes, not months.
• Decide: persist, pivot, or stop, against criteria you set up front instead of arguing about under pressure.

It is AI-agnostic on purpose. Whether the team is doing classical ML, RAG, or agents, the framing skill is the same, and so are the ways it goes wrong.

Format

A half-day working session, about 3.5 hours, run remotely. One team, one focused pass through your projects. The ideal group is five to ten people who frame and scope AI work: ML and AI engineers, the PMs who write the specs, and the leads accountable for the outcome. Mixed seniority helps rather than hurts, because framing arguments are usually cross-level.

Billing runs through Maven, so procurement is straightforward.

Pre-work

Before we meet I ask for two things: a short intake call to tailor the examples to your domain, and two or three real projects from the team. The workshop is only as good as the projects on the table.

About the instructor

Rajiv Shah is an AI Agentic Engineer at OpenHands. He has worked hands-on across 100+ AI use cases in enterprise, startup, and research, including enterprise AI at State Farm and Caterpillar and work with Hugging Face, Snowflake, and Snorkel. The failures taught him more than the wins, and they almost always came down to framing, not algorithms. That pattern is documented in a database of 250+ production AI reframes that backs every example in the workshop.

Set up a call

A 20-minute call tells us whether your team's projects are a fit. Get in touch →