A2X helps technical teams move from casual Claude use to a real 10x workflow. Slides, recordings, cheat sheets, and the tools we built together — all here, ready when you need them.
A2X is a one-person practice focused on a single promise: leave every workshop with skills your team can use on real work — and a place to come back when you forget the details.
I'm Tomer Harari Hamam, PhD — an AI engineer, founder, and consultant. I hold a PhD in Electrical and Computer Engineering from Georgia Tech, and I've spent more than a decade building real-world AI — leading motion-planning and decision-making for Level-4 autonomous vehicles at Imagry, and co-founding a deep-tech startup that builds autonomous AI agents for industry.
I use Claude Code every day — building MCP servers and custom skills, and publishing my own plugin marketplace — and I've turned that hands-on practice into a workshop methodology for technical teams. A2X is where I help organizations adopt Claude and Claude Code with the same rigor I bring to my own work: practical, customized, and built to outlast the session.
Live two-session workshops tailored to your codebase and the way your team actually works — not a generic AI talk.
The methods come from daily practice: MCP servers, custom skills, hooks, and a published Claude Code marketplace — battle-tested before they reach your team.
Every workshop comes with a home base — slides, cheat sheets, interactive explainers, and office hours — so what you learned actually sticks.
The goal is not another AI talk. It is a practical adoption sprint: make your team faster, safer, and more consistent on real engineering work.
Best next step: send a short note about team size, current Claude usage, your biggest bottleneck, and timing. Tomer will reply with the right fit-check path.
The reference library for workshop alumni. Start with the essentials below, open the live resource hub, or inspect the public GitHub source for the full workshop bundle.
The single most useful mental model: Claude Code is a harness running a model in a tool-using loop — and almost every problem is a harness problem you control.
Read the guide →Four rules to paste into your CLAUDE.md — Think, Simple, Surgical, Goal-driven — each killing a common failure mode, with before/after examples.
Open cheat sheet →An interactive, do-it-yourself walkthrough from "what's a skill?" to a working Claude Code skill of your own.
Start building →The public GitHub repo behind the hub: templates, tutorials, sanitized demos, usage maps, and public-safe handouts from the Claude Code workshop track.
View source on GitHub →Additional Claude Code explainers and workshop handouts. Published previews open directly; still-unfinished items remain marked as coming soon.
Which Claude subscription or API path actually unlocks Claude Code, and why Claude.ai and API billing are separate pools.
Open preview →Why an interactive HTML mockup beats a Markdown to-do list as the artifact you react to before any code is written.
Open preview →Verification loops, attention as the real bottleneck, and running work in the background with /loop.
Open preview →Memory stores explained — scopes, read-write vs read-only, the steering prompt — plus "Dreaming", the async cleanup that keeps memory fresh.
Open preview →Turn Claude from a tool into a teammate: routines built from trigger × context × steerability, so work starts without you typing.
Open preview →Moving Claude Code work across desktop, web, remote, and phone — and matching the surface to the task.
Open preview →How hooks fire automatically at fixed points in the loop to enforce guardrails without spending your attention.
Open preview →A standalone model-literacy explainer on how modern open-weight models work.
Open preview →A one-page, printable cheat sheet for prompting Claude and Claude Code well.
Coming soonA standing place to bring questions after the workshop — recurring office hours and an alumni channel.
Coming soonA focused, two-session Claude Code workshop — customized to your codebase and the way your team already works — so the speedup actually lands and stays. The first step is simple: send Tomer a short note about your team, current Claude usage, biggest bottleneck, and timing.
Send the practical details by email—team size, current Claude usage, main bottleneck, and timing—or choose a 20-minute fit-check time directly. Tomer will reply with the recommended next step.