Prompt Advisers
82.2k
YouTube subscribers
5
O'Reilly live courses
1.3k
community members
500+
businesses advised
The program
No fixed calendar.
Scoped to your company.
Done when your team ships.
Every engagement moves through the same four phases. Some teams run them in a few intensive weeks, others across a quarter. The program ends when shipping with AI is muscle memory, not on a date we picked in advance.
II / The arc
Claude Code and Codex as an operating system, not a chatbot. Everyone gets a workspace, the mental model, and a first real win before the first session ends.
III / Formats
Executive briefing
A half-day for leadership on what AI changes, what it does not, and where to place the first bets.
Team bootcamp
The full program. Technical and non-technical tracks in one shared workspace, shipping real work at your company's pace.
Non-technical teams
Operations, marketing, and finance learn to delegate real tasks to Claude Code and Codex.
Train the trainer
We leave an internal champion who can onboard every future hire without us.
IV / Fair game
If your team touches Claude, it’s fair game.
Sixty-plus topics across eight tracks, from first prompts to production agent systems. Everything on the menu can shape your curriculum. Or take a track we already run.
AI fluency for everyone
For every employee. No code required.
- What Claude can do
- Spotting AI mistakes
- Three-part prompts
- When to use AI, when not to
- Data privacy basics
Prompt engineering
Getting exactly what you want out of the model.
- Few-shot examples
- Chain-of-thought reasoning
- Fast prompt iteration
- Structured outputs
- Long-document handling
- Testing prompt quality
- Prompt caching
Claude Code fundamentals
The operating system, not the chatbot.
- Setup across IDE, CLI, and web
- File operations
- Artifacts and live preview
- The permission model
- Shell and terminal automation
- Git without breaking context
- Verify before deploy
- CLAUDE.md project config
- Settings and customization
Agents and automation
Work that runs while your team does something else.
- Agent fundamentals
- Skills, built-in and custom
- Hooks and event triggers
- Custom tools
- Connecting MCP servers
- Headless Claude in CI/CD
- Subagents and delegation
- Agent memory
- Failure handling and retries
Building on the API
For teams shipping Claude inside their own product.
- The Messages API
- Tool calling
- Vision and multimodal
- Streaming responses
- Token budgeting
- Cost optimization
- Prompt caching at scale
- Monitoring in production
- Rate limits and scale
Advanced agent systems
Production-grade autonomy with guardrails.
- Plan mode and approval gates
- Multi-agent orchestration
- Database and system integration
- Persistent state
- Building MCP servers
- The Agent SDK
- Performance tuning
- Sandboxing and isolation
Enterprise rollout
Taking it across an organization safely.
- Enterprise and Teams setup
- Roles and access control
- Shared projects
- Cost tracking and chargeback
- Audit and compliance
- Security hardening
- Proving ROI
- Adoption and change management
Claude everywhere
Every surface Claude lives on.
- Claude on the web
- The desktop app
- The command line
- IDE extensions
- Claude in Slack
- Claude in Chrome
- Computer use
Or take a track we already run
10 topics · reads like a half-day session
AI fluency for the whole company
A half-day primer for non-technical teams on what Claude does, prompt basics, and responsible use.
13 topics · reads like a multi-day bootcamp
Claude Code bootcamp for engineers
Hands-on intensive. File operations, git, automation, and verification on your real codebase.
10 topics · reads like a half-day session
AI strategy for leadership
An executive briefing on capabilities, governance, security, and where to place the first bets.
16 topics · reads like a multi-day bootcamp
Agent engineering masterclass
Autonomous agents, MCP servers, multi-agent systems, and production deployment patterns.
13 topics · reads like a multi-day bootcamp
Building products on Claude
The API end to end. Tool use, vision, streaming, cost, and scale for teams shipping AI features.
Take one as-is, remix the menu, or hand us an outcome.
V / On the public record
The same material we teach privately runs as live courses on O’Reilly, in front of thousands of learners. Mark Kashef holds a master’s in AI from Queen’s University and has advised hundreds of businesses; your team learns from him directly, not a junior facilitator with a deck.
What clients say
“To say Mark went "above and beyond" is an absolute understatement. I got way more than I paid for. I have some background with prompt engineering, but Mark blew my mind. He's incredibly knowledgeable in what he does.”
Verified client review
“Mark did an amazing job and went well beyond what I asked for. He's not only brilliant, he wants to set you up for success, so he explains what he's doing and how you can replicate the results.”
Kathy · Verified client review
“A top tier expert at prompt engineering. He not only provides you with well thought out prompts, he takes the time to explain the prompts and teaches you how to apply them.”
Verified client review
VI / Questions
Or send a note