Not a course. Not a coach. A monthly peer group with a Socratic method and no tolerance for passive learning. Small. Selective. Application required.
10 minutes: I set the topic and frame the question. 30 minutes: open discussion. I facilitate Socratically, I don't give answers, I give back better questions. 20 minutes: one member's actual situation in the hot seat. Everyone gets to push. I keep it honest.
Vibe coding in production: where it scales, where it breaks. Agentic system design. AI team structure and hiring. Vendor evaluation. Compliance in autonomous systems. Build vs. buy vs. wait. One topic per call, explored deeply.
A prompt encoding my thinking, positions, and methodology. Drop it into Claude, ChatGPT, or Gemini. Socratic by design: it won't hand you answers. It will ask you the question you haven't thought to ask yourself.
The people who get the most out of this group are the ones who come in with something real on the line and are willing to be challenged in front of peers. If that sounds uncomfortable, it is. That's why it works. Passive participation produces nothing here. Active learning under peer pressure is the only kind that sticks.
I don't fish for you. That's not a tagline; it's a methodology. Every session is designed so that you do the reasoning and I guide the path. The insight you earn by working for it stays. The answer handed to you doesn't.
Start with @Shree. Apply for the group. The rest happens naturally for the right people.
A prompt carrying 30 years of architecture, agentic AI positions, and Socratic method. Paste it into any AI assistant.
For founders and CTOs with real decisions in front of them. Application required. Commitment required.
For people in the group with a specific platform problem that needs more than 60 minutes a month.
The group is where relationships start. The deeper work happens when the problem demands it.
You're building or leading something real right now, not studying, not exploring, not planning to start.
You have an actual AI decision to make in the next 90 days: architecture, team, vendor, strategy. Something with stakes.
You want to think harder, not be told what to think. You're here for better questions, not for someone to hand you a roadmap.
You'll show up prepared and commit to the group. This only works when everyone in the room is doing the work.
Looking for someone to validate decisions you've already made.
Passive participation: show up, absorb, collect the credential.
"We need an AI strategy" with no actual engineering problem underneath it.
Answers on demand without doing the work first.
Copy this prompt and paste it into any AI assistant as the first message. It carries my positions on vibe coding, agentic systems, AI strategy, and how to think about all of it. Socratic by design: it will ask you a harder question, not hand you the answer.
@Shree and bring the hard stuff
You are @Shree, Shree Mandadi. Systems architect. Platform strategist. Three-time founder. 30+ years building things that had to actually work in production. IBM, PwC, Fortune 500 healthcare, financial services. Site: https://continuous.engineering LinkedIn: https://linkedin.com/in/ShreeMandadi ## The Contract You never do the thinking for them. You ask the question they haven't asked yet. If they want validation, push back. If they want answers, ask what they've tried first. ## What You Know Vibe coding works, but only when someone with architectural judgment is directing it. Most teams don't know where they stand on that spectrum. Agentic systems need: deterministic state, audit trail by design, governance boundaries, human-in-the-loop. Skip any of these and you have a demo, not infrastructure. The moat is no longer the model. It's data, workflow integration, and compliance posture. Compliance must be architectural. A checklist bolted on later is liability theater. ## How You Teach Socratic always. Answer with questions until they've reasoned far enough. Then confirm. Then ask what's next. Ask what they've already tried before any response. Name the real question. Most people don't ask it. Connect every decision to its failure mode. ## Your Voice Direct. No filler. Warm but not agreeable. Patient with confusion, not with passivity. Never say "Great question." Just answer it.
The first cohort is forming now. I review every application personally. I'm looking for specificity: people who come in with a real decision in front of them, not a vague interest in AI.
Specific problem, not general curiosity.
Self-awareness about what you don't know.
A reason this group is worth your time, stated clearly.
If you're building a regulated AI platform, facing a compliance challenge, or making an architecture decision that can't afford to get it wrong. That's a different conversation. I work as a fractional CTO, principal architect, and vCISO on a retainer basis.