Track I
Foundations
A clear-eyed mental model for how modern models work — no mystique, no hype, just a working map of the landscape.
How Modern Models Work
What LLMs and multimodal models actually do under the hood, and what that means for what they can and can't do for you.
The Provider Landscape
Cost, latency, and quality trade-offs across OpenAI, Anthropic, Google, and the open-source ecosystem.
Strengths & Failure Modes
Where today's models shine, where they break, and how to design around the failure modes that matter.
Vocabulary & Concepts
Tokens, context, embeddings, retrieval, tool use — the shared language you need to make decisions with your team.
Track II
Prompting, Evals & Agents
Prompt design as an engineering discipline — with versions, tests, and reviews. Plus the patterns for composing agents that plan, act, and self-correct.
Structured Prompting
Patterns for prompts that are robust, testable, and easy to maintain across model upgrades.
Retrieval & Context
RAG, context management, and the supporting infrastructure that lets models reason over your data.
Evaluations
Writing evals so you can tell when a change actually makes things better instead of just feeling better.
Agentic Patterns
Composing multiple models and tools into agents — coding agents, research agents, back-office automation — with real guardrails.
Track III
AI in the Workplace
Patterns for embedding AI into the real work of a company — research, writing, support, sales, ops, and engineering — with the human review loops that keep output trustworthy at scale.
Operations & Back-Office
Automating the high-volume, low-judgment work that quietly drains your team's hours.
Sales & Support
AI that augments customer-facing teams without making the experience feel automated.
Research & Writing
Workflows that turn AI into a research partner and editor — not a content firehose.
Engineering Productivity
Coding agents, review assistants, and the cultural shifts needed to actually capture the productivity gains.
For Leaders
AI strategy for founders, executives, and team leads
How to set an AI strategy, prioritize use cases, organize teams around AI, and avoid the common traps of pilots that never ship.
- Picking the right first use cases — high leverage, low risk, measurable.
- Make-vs-buy decisions and build order across a multi-quarter roadmap.
- Organizing teams: where AI lives, who owns it, how it's reviewed.
- Measuring real ROI instead of vanity adoption metrics.
- Data, privacy, and governance bars your customers and regulators expect.
Emergent Capabilities
Tracking the frontier together
Frontier Tracking
Every new release unlocks new capabilities — we run structured experiments and translate them into product opportunities.
Long Context & Multimodal
Working with very long context, native multimodality, and reasoning models — and knowing when each actually helps.
Reasoning & Tool Use
How agentic and reasoning models change what's possible in research, analysis, and complex workflows.
Hands-On
Labs, clinics, and shipped systems
Build Labs
Hands-on labs where you build real workflows on your own data — not toy demos.
Office-Hours Clinics
Mentors review what you've shipped, debug what's stuck, and pressure-test what you're planning next.
Peer Network
A community of operators solving similar problems — the ongoing network is half the value.