Skip to content

wjlgatech/omega-founders

Repository files navigation

OmegaFounders Logo

OmegaFounders — Silicon Valley AI Cohort

The AI job market is splitting in two. One side has 3.2 open roles per qualified candidate. The other side is getting automated.

OmegaFounders is a 12-day Silicon Valley cohort for founders and engineers who want to land on the right side of that split — by shipping real AI systems, earning the Claude Certified Architect credential, and demoing in front of investors.

"The desert shall rejoice and blossom." — Isaiah 35


What This Repo Is

Path What's in it
landing/ Landing pages — dark premium & Golden Hour variants
docs/10-DAY-PLAN.md Full 12-day sprint plan (days 1–10 learn/build/post, 11 demo, 12 graduation)
docs/STUDY-RESOURCES.md Certification prep resources
skills/claude-architect-mastery/SKILL.md AI-powered exam prep skill (works inside Claude Desktop)
learn-build-teach/ Daily public learning logs — every cohort member's daily entry
.github/ PR templates, issue templates, CI workflows

Why Join

📈 The K-Shaped Market — Why timing matters right now

Real-life benefit: The AI job market is not rising uniformly. It is splitting into a K — top track accelerates, bottom track contracts. The inflection point is whether you can build and evaluate AI systems, not just use them. The K-split creates 3.2 open roles for every qualified AI architect candidate, right now, in 2026.

The innovation: Most AI "education" teaches prompt tricks. OmegaFounders teaches the 7 core professional skills that show up in every job description for AI Architect, AI Product Lead, and AI Engineer roles: Specification Precision, Evaluation & Quality Judgment, Task Decomposition, Failure Pattern Recognition, Trust & Security Design, Context Architecture, and Cost & Token Economics.

How it works: The 12-day sprint is structured as a Learn → Build → Teach flywheel. Each day maps to a specific certification domain, produces a shipped artifact, and requires a public post. By Day 12 you have a portfolio, a credential, and a demo — not just course completion.

📋 See the full 12-day plan →

🤖 Claude Architect Mastery Skill — AI that teaches you AI

Real-life benefit: Exam prep is usually passive (read → forget). This skill is active: it quizzes you in exam format, simulates the 6 real scenarios, tracks your mastery across domains, and coaches your weakest areas. It turns any Claude Desktop session into a personalized certification tutor.

The innovation: The skill encodes practitioner-level understanding, not just exam trivia. It knows the why behind every pattern — why hooks beat prompts for compliance, why 4–5 tools is optimal per agent, why the same model can't review its own output. It also drills the 6 AI failure modes that separate architects from prompt engineers.

How it works: Drop skills/claude-architect-mastery/SKILL.md into .claude/skills/ in your Claude Desktop workspace. Then invoke it with: quiz me on Domain 1, simulate Scenario 3, explain failure modes, show my mastery, or Day 4 study session. It runs in a forked context so it doesn't pollute your main conversation.

🔧 View the skill →

🏗️ 12-Day Build Structure — Ship every day, not just at the end

Real-life benefit: Most bootcamps front-load learning and back-load building. You come out with knowledge but no portfolio. OmegaFounders structures every day as: 45 min Learn → 90 min Build → 30 min Teach (public post). By Day 10 you have 10 shipped artifacts and 10 public posts. By Day 12 you have a demo and a credential.

The innovation: The 10-day learning track is mapped 1:1 to the Claude Certified Architect exam domains, weighted by exam percentage (D1: 27%, D3: 20%, D4: 20%, D2: 18%, D5: 15%). Day 11 is Founder Demo Day. Day 12 is Invited Speakers + Graduation. The structure forces compound learning — each day builds on the last.

How it works: Each day in docs/10-DAY-PLAN.md has: task statement numbers, real-world skill built, active failure modes to recognize, and a concrete BUILD artifact. The plan integrates the 7 core professional skills and 6 failure modes across all 10 learning days — not as separate chapters, but as the operating system for each day's work.

📅 Full sprint plan →

🌐 Landing Pages — Two variants, zero JavaScript frameworks

Real-life benefit: The landing page converts visitors into cohort applicants. Two design variants let you A/B test conversion: dark premium (elite Silicon Valley vibe) vs. Golden Hour (warm community vibe). Both are single-file HTML — no build step, no npm, deployable anywhere.

The innovation: Both pages use Apple-grade micro-interactions: backdrop-filter: saturate(180%) blur(20px) frosted glass nav, meshDrift animated gradient mesh, fadeSlideUp scroll reveals via IntersectionObserver, and a CSS-only ticker tape. The dark version adds a live K-shaped market visualization with the 7 skills grid. Neither page has a framework dependency.

How it works: Open landing/index.html or landing/golden-hour.html in any browser. No build step. Drop either onto any static host (Netlify, Vercel, GitHub Pages, Cloudflare Pages) and it works. The K-shaped market section is interactive and data-driven from CSS variables — change --gold in 1 line to rebrand the entire page.

🌑 Dark version → · 🌅 Golden Hour →

🤝 Multi-Contributor Flywheel — Each cohort member's daily log lives here

Real-life benefit: Public learning creates accountability and audience. When 20 founders are each posting daily logs to this repo, the repo itself becomes a content marketing machine. Each PR is a signal of activity; each daily log is a tweet thread waiting to happen.

The innovation: The learn-build-teach/ directory has a standard daily log template. Each contributor creates learn-build-teach/[github-handle]/day-[N].md. The PR template enforces the structure: what you learned, what you built, where you posted. This is a living portfolio, not a graveyard of completed coursework.

How it works: Fork the repo → create your folder → submit a PR per day using the daily log template. CI validates your markdown has the required sections. See CONTRIBUTING.md for the full workflow.

📝 Daily log template → · 🤝 Contributor guide →


Quick Start (Cohort Member)

# 1. Fork this repo to your account
# 2. Clone your fork
git clone https://github.com/YOUR_HANDLE/omega-founders.git
cd omega-founders

# 3. Set up the exam prep skill
cp skills/claude-architect-mastery/SKILL.md ~/.claude/skills/

# 4. Create your daily log folder
mkdir -p learn-build-teach/YOUR_HANDLE
cp learn-build-teach/TEMPLATE.md learn-build-teach/YOUR_HANDLE/day-01.md

# 5. Read the 12-day plan
open docs/10-DAY-PLAN.md

# 6. Start Day 1
# Then submit a PR with your day-01.md at end of day

Quick Start (Contributor / Builder)

# Clone the main repo
git clone https://github.com/wjlgatech/omega-founders.git
cd omega-founders

# Create a feature branch (one branch per PR)
git checkout -b feat/your-feature-name

# Make your changes
# Then open a PR using the PR template

See CONTRIBUTING.md for branch naming, commit conventions, and the PR checklist.


Cohort Schedule

Day Focus Domain
1 Agentic Loops D1 (27%)
2 Multi-Agent Architecture D1
3 Tool Design D2 (18%)
4 MCP Integration D2
5 Claude Code + CLAUDE.md D3 (20%)
6 Workflows & CI/CD D3
7 Prompt Engineering D4 (20%)
8 Structured Output & Batch D4
9 Context & Reliability D5 (15%)
10 Exam + Final Ship All
11 Founder Demo Day 🎯 Ship
12 Invited Speakers + Graduation 🏆 Celebrate

The 7 Core AI Professional Skills

Every day in the cohort develops at least one of these:

# Skill Exam Domain
1 Specification Precision D4
2 Evaluation & Quality Judgment D4, D5
3 Task Decomposition & Delegation D1
4 Failure Pattern Recognition D5, D1
5 Trust & Security Design D1, D2
6 Context Architecture D5, D3
7 Cost & Token Economics D4, D5

The 6 AI Failure Modes

Know these cold. They show up in the exam and in every production system:

# Failure Mode Why It Matters
1 Context Degradation Quality drops silently as conversation grows
2 Specification Drift Output drifts from intent across agent handoffs
3 Sycophantic Confirmation Model agrees with wrong premises
4 Tool Selection Errors Wrong tool called, or no tool when needed
5 Cascading Failure Error amplifies through multi-agent pipeline
6 Silent Failure ⚠️ System "succeeds" but produces wrong results

Full deep-dive on all 6 →


Star History

If this repo is helping you — star it. Every star is a vote that the right side of the K-split should be accessible, not gatekept.


License

MIT. Build on it, ship with it, teach from it.


Built during a 12-day sprint. Powered by stubbornness, Claude, and Isaiah 35.

About

OmegaFounders — 12-day Silicon Valley AI Architect Cohort. Build, certify, and ship agentic AI with Claude.

Topics

Resources

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors