Document type: Strategic framework + technical spec
Owner: Harnoor Minhas / Cloud 8 Data
Date: 2026-04-27
Status: Pre-launch — Cohort 1
Builds on: live-demo-script.md · parent-packet.md · stripe-products.md
---
Students and parents should feel they are in Harnoor's classroom. Third-party platforms — AWS Skill Builder, Replit, Hugging Face, Colab — are tools Harnoor hands them, not brands they are handed off to. The frame never breaks.
1. Branded shell — academy.cloud8data.com
Every link students receive, every email they open, every landing page they hit — the URL is academy.cloud8data.com. Behind this you can iframe or redirect to external platforms, but the parent URL never changes until the student is ready to create their own account on a specific tool.
Set up academy.cloud8data.com as a subdomain pointing to the TITAN /academy route. The nav header reads "Cloud 8 Academy" with Harnoor's photo and name in the top-left, not a third-party logo.
2. The parent-packet language model
In all parent-facing and student-facing copy, the framing is always: "In this module, Harnoor is having you use [tool] because it gives you hands-on time with a real [AWS / AI / Python] environment. Think of it as your lab workbench." The platform is the workbench; Harnoor is the teacher who assigned it.
Avoid phrases like "Go to AWS Educate" as a standalone instruction. Always lead with: "Harnoor's Week 9 lab assignment uses AWS Educate to give you real cloud console access. Here's your setup link."
3. Iframe wrapping for platform pages where possible
Google Colab notebooks can be embedded or linked with pre-loaded notebooks (using the colab.research.google.com/github/[repo]/blob/main/[notebook].ipynb URL pattern). Hugging Face Spaces are embeddable via iframe (<iframe src="https://huggingface.co/spaces/[org]/[space]">). Replit Repls can be embedded as iframes.
For platforms that cannot be iframed (AWS Skill Builder enforces X-Frame-Options), the handoff page on academy.cloud8data.com shows a full-bleed branded card: Harnoor's photo, the assignment context, the learning objective, and then a single CTA button "Open Your AWS Lab" that opens in a new tab. The student never feels dropped.
4. Lab completion returns to TITAN
After completing an external lab, the student returns to academy.cloud8data.com to mark it done (screenshot upload + "I completed this" checkbox). The completion loop closes inside your brand, not inside AWS or Replit.
5. Email and Discord — all Harnoor's voice
All automated emails come from harnoor@cloud8data.com (or academy@cloud8data.com if volume-separated). Weekly progress emails to parents are signed by Harnoor personally. The Discord server is named "Cloud 8 Academy" with Harnoor's avatar as the server icon.
6. Certification — your brand, not theirs
The Certificate of Completion is issued on Cloud 8 Data / HM Tech Solutions LLC letterhead, signed by Harnoor, co-branded with the AWS Partner logo (which you are licensed to use as an AWS Partner). No third-party platform logo appears on the certificate. When students mention their credential on LinkedIn, they say "Cloud 8 Data (AWS Partner)" — that is the brand that carries weight, not "I completed AWS Educate."
---
URL: https://aws.amazon.com/education/awseducate/
Free for students? Yes — entirely free. No credit card required.
Age floor: 14+ globally (16+ in EU countries; 18+ in Algeria, Lebanon, Portugal).
Welcome credits: $30–$50 in AWS Promotional Credits on enrollment.
What's covered: 600+ hours of cloud courses; hands-on missions earning AWS Educate Badges (micro-credentials); AI services intro (Amazon Lex, Polly, Rekognition); Cloud Literacy Badge pathway for absolute beginners; Cloud Pathways for intermediate learners with live AWS console access.
Sign-up process for under-18: Email-only signup — no school .edu email required, no credit card. Students provide DOB at registration; the platform gates 14+ content. Parental consent form recommended but not technically enforced by the platform.
Branding: AWS-branded entirely. Use the parent-packet frame above to keep Cloud 8 forward.
Best for Jumpstart: Weeks 9–12. Pairs with Cloud Quest (below) for the game-based entry point.
URL: https://skillbuilder.aws/ and https://skillbuilder.aws/cloudquest
Free for students? AWS Cloud Quest: Cloud Practitioner role is FREE on the free Skill Builder tier. Other Cloud Quest roles (Solutions Architect, Generative AI, Security) require a paid subscription (~$29/month individual).
What it is: A role-playing game where students build real AWS infrastructure while solving in-game quests. Cloud Practitioner role = foundational cloud concepts, EC2, S3, IAM in a live AWS sandbox. Genuinely engaging for high schoolers — it looks and feels like a game, not a course.
Age floor: No explicit minimum stated; AWS terms of service require 13+ for account creation.
Sign-up process: Free AWS account required (email, no credit card for the free tier labs within Skill Builder).
Branding: AWS-branded. Cannot be iframed. Use the handoff-card pattern above.
Best for Jumpstart: Week 9 opener — this is the "wow" moment where students realize they're touching real AWS. Budget 90 minutes for the first Cloud Quest session.
IMPORTANT NOTE: QwikLabs is NOT AWS's product. QwikLabs became Google Cloud Skills Boost in 2022. AWS's equivalent is AWS Skill Builder / Cloud Quest. When Harnoor referenced "AWS QwikLabs," the correct current name is AWS Skill Builder with Cloud Quest.
URL: https://cloudskillsboost.google/
Free for students? GEAR program offers 35 monthly credits for free; the Google Arcade challenge (free seasonal program) provides additional credits. College students can request 200 credits with institution verification. High school students without a school-issued Google account may need to self-register with the GEAR program.
What it is: The platform formerly known as QwikLabs (rebranded 2022). Browser-based labs in a live GCP environment. Strong data engineering and AI/ML track content (BigQuery, Vertex AI, Looker Studio).
Age floor: Google account required; standard Google account TOS is 13+.
Branding: Google-branded. Cannot be iframed. Same handoff-card approach as AWS.
Best for Jumpstart: Optional supplement in Weeks 5–8 for students who want to explore GCP alongside AWS. Not core to the Jumpstart program — reduces complexity to keep it AWS-first.
URL: https://learn.microsoft.com/
Free for students? The inline Microsoft Learn sandboxes (the embedded Azure playgrounds within learning modules) are technically available but have been experiencing access issues as of 2025-2026. Multiple forum reports indicate the sandbox activation is inconsistent. Azure for Students ($100 free credits) requires a verified school email (.edu) and an age-13+ account.
Verdict for Jumpstart: Do NOT make Microsoft Learn sandboxes a primary dependency. The inconsistent availability makes it unreliable for a structured cohort. Use Azure for Students as a bonus resource for students who want to explore multi-cloud, not as a core lab environment.
URL: https://www.codecademy.com/student-center
Free for students? Not free — student discount is 35%+ off Pro. Regular Pro is ~$19.99/month; with student discount approximately $12–$13/month. Verification requires student ID or enrollment proof.
What it is: Structured Python, data science, and AI/ML learning paths with in-browser code editor. No external setup required.
Age floor: 13+ with parental consent for under-18.
Verdict for Jumpstart: Good supplemental resource for students who want extra structured practice, but do not make it a required purchase — it adds cost friction for families. Use Replit (free) as the primary hands-on environment instead.
URL: https://replit.com/partners/education/students
Free for students? Yes — free Starter plan includes browser-based coding in 100+ languages, AI assist (limited), real-time collaboration, and the ability to fork templates instantly. Teachers can create Teams for free (public schools free; 80% discount otherwise), which gives private workspaces, auto-forking of starter code for each student, and a gradebook.
What it is: Browser-based IDE — no installation, no environment setup. Students open a URL and start coding in Python within 30 seconds. Perfect for Week 1 when students have zero setup.
Age floor: 13+ account; under-13 requires additional consent (COPPA). Parental consent built into the Teams for Education flow.
Starter template for Jumpstart: Groq Chatbot Quickstart — https://replit.com/@replit-matt/Chat-with-Groq?v=1#main.py. Students fork this, get a free Groq API key at console.groq.com, and have a working AI chatbot in under 20 minutes. This is your demo class lab (see Section J).
Branding: Replit-branded but the Repl URL can be customized. The completion loop returns to TITAN academy.
Best for Jumpstart: Weeks 1–8, primary coding environment. Zero friction for students with no prior setup.
URL: https://education.github.com/
Free for students? Yes — GitHub Student Developer Pack is free for verified students 13+. Includes: GitHub Pro (free), GitHub Codespaces (180 core-hours/month free — enough for a full semester of coursework), and 70+ partner tool credits including Microsoft Azure ($100 for 18+, free services for 13–17).
Verification requirement: Students must be enrolled at an "accredited institution." For high schoolers this means their high school must be accredited (virtually all are). They submit proof via their school email or an enrollment document scan.
What Codespaces gives you: A VS Code environment running in the browser backed by a cloud VM. Students can run Jupyter notebooks, Python scripts, and full dev environments with zero local setup. Heavier than Replit but more production-like.
Best for Jumpstart: Week 13–16 Capstone. Students graduate from Replit (beginner) to Codespaces + GitHub repositories (production-ready). The capstone project lives in a public GitHub repo — this is the college-application artifact.
Branding: GitHub-branded, but the student's public GitHub profile becomes their portfolio. Frame it as: "Your GitHub profile is your tech resume. We are building it together."
URL: https://huggingface.co/spaces
Free for students? Yes — free account, no credit card. CPU Basic hosting for Spaces is free indefinitely. GPU instances are paid ($0.40/hour+). For student AI demos (Gradio or Streamlit front-ends on small models), CPU Basic is sufficient.
What it is: Serverless hosting for AI demos. Students build a Gradio or Streamlit app locally (or in Colab/Codespaces), push to Hugging Face, and get a public URL they can share with anyone. The Space link is the capstone demo URL on LinkedIn and college applications.
Education program: Hugging Face for Education (https://huggingface.co/blog/education) provides free classroom infrastructure, pre-built course materials, and the ability to create a class organization on the Hub where student submissions are private to the cohort.
Age floor: Standard account is 13+; no parental-specific consent flow beyond standard TOS.
Branding: Hugging Face-branded public URLs, but students own their Spaces. Frame as: "Your Hugging Face Space is your public AI portfolio."
Best for Jumpstart: Weeks 13–16 Capstone deployment. Every student ships their capstone as a live HF Space.
URL: https://colab.research.google.com/
Free for students? Yes — free with a Google account. Free tier provides CPU runtime + limited GPU access. Sessions disconnect after 90 minutes of inactivity; max 12-hour runtimes per session.
What it is: Jupyter notebooks running in the cloud. No installation. Students open a .ipynb link and run Python immediately. Pre-loaded datasets via Google Drive integration.
Best for Jumpstart: Weeks 5–8, LLM and data labs. Harnoor can pre-build notebooks hosted on GitHub and share Colab "Open in Colab" links: https://colab.research.google.com/github/[repo]/blob/main/[notebook].ipynb. Students fork and run — zero setup.
Limitation to flag: Free tier GPU is not guaranteed. For the Jumpstart curriculum, all CPU-runnable notebooks (pandas, sklearn, basic HuggingFace inference) — no GPU dependency needed until capstone, at which point GitHub Codespaces or HF Spaces handles it.
Branding: Google-branded. Negligible — students are focused on the notebook content (which you author), not the platform.
URL: https://jupyterlite.readthedocs.io/
Free for students? Yes — runs entirely in the browser, no server required. Powered by WebAssembly (Pyodide).
What it is: Jupyter notebooks running client-side in the browser. Nothing to install, no account required, no sign-up friction. Supports Python (via Pyodide), matplotlib, pandas for basic data work.
Limitation: No network access from kernel, so cannot call external APIs. Good for self-contained data exercises; not suitable for LLM API labs.
Best for Jumpstart: Week 1–2 Python exercises where you want zero sign-up friction for an intro taste. Not suitable for Weeks 5+ where API access is needed.
Branding: Invisible — you can host a JupyterLite instance on academy.cloud8data.com/lab using GitHub Pages or TITAN's static hosting.
URL: https://claude.com/solutions/education
Free for students? No standard student discount or free tier as of April 2026. Anthropic's Claude for Education is a campus-licensing program negotiated institution-to-institution (universities, not high schools). No individual high-school student offer exists yet.
What is available: (a) Anthropic Campus Ambassador program — university students, not high school. (b) Builder Club API credits at select universities. (c) Small free API credits ($5) on new API account creation. (d) claude.ai free tier (limited message cap) for general use.
Verdict for Jumpstart: Do not position Anthropic as a platform students "use" for labs. Instead, frame Claude as the professional AI tool Harnoor demonstrates in the live class (claude.ai free tier is sufficient for the demo). Students don't need an API key for Claude during the program — Groq (free API, Llama-3) handles the chatbot lab, and Hugging Face handles model inference.
---
This is the concrete spec for the TITAN /academy route. Each student logs in and sees their personalized tile board.
Header: Cloud 8 Academy | [Student Name] | Week [X] of 16 | [Overall Progress Bar]
[TODAY tile — pinned top]
Row 1: [Video tile] [Video tile] [Assignment tile]
Row 2: [Lab tile] [Quiz tile] [Lab tile]
...
[Topic Progress Bars at bottom]
{
"type": "video",
"title": "Intro to Pandas DataFrames",
"youtube_id": "vmEHCJofslg",
"eta_minutes": 18,
"status": "todo | in-progress | done",
"completion_method": "checkbox",
"icon": "▶"
}
{
"type": "assignment",
"title": "Week 3 Reflection — Your first data analysis",
"description": "In 150–200 words, describe what surprised you most about the Pandas exercise. What would you have done differently?",
"due_date": "2026-09-15",
"submit_method": "text-response | link",
"status": "todo | submitted | reviewed",
"icon": "✏"
}
{
"type": "lab",
"title": "AWS Cloud Quest — Mission 1: Cloud Practitioner",
"platform": "AWS Skill Builder",
"external_url": "https://skillbuilder.aws/cloudquest",
"setup_note": "You'll need a free AWS account. Instructions in the Lab Setup guide.",
"eta_minutes": 90,
"completion_method": "screenshot-upload + checkbox",
"status": "todo | in-progress | done",
"icon": "⚗"
}
{
"type": "quiz",
"title": "Week 1 Knowledge Check — Python Basics",
"questions": 5,
"format": "multiple-choice",
"passing_score": 70,
"attempts_allowed": 3,
"status": "todo | passed | failed | expired",
"icon": "?"
}
One bar per topic module:
Each bar = (completed tiles / total tiles) × 100. Color: green ≥ 80%, yellow 50–79%, red < 50%.
All tiles share these base fields:
{
"id": "w03-lab-01",
"week": 3,
"module": "python-basics",
"type": "lab | video | assignment | quiz",
"title": "string",
"eta_minutes": 45,
"status": "todo | in-progress | done | submitted | passed | failed",
"required": true,
"visible_after": "w03-video-01"
}
---
Cost: Free (personal), $8/month (Plus for more blocks)
What it does: Page hierarchy = course > module > lesson. Students can be added as guests. No progress tracking, no quiz engine, no completion checkboxes.
Verdict: Usable for content delivery but zero tracking. Requires Harnoor to manually check progress. Not recommended as primary — fine as a knowledge base supplement.
Cost: Starter $29/month (annual) with 7.5% transaction fee; Builder $69/month with 0% fee.
For 30 students: At Builder tier, $69/month flat regardless of enrollment. Transaction fee gone. Decent video hosting, course content, drip scheduling.
Limitations: No community features. No hands-on lab integration. Feels like passive video watching, not active learning. Certificate generation is basic.
Verdict: Overkill at this stage. You are already using Stripe for payments — Teachable adds redundant payment infrastructure and costs $828/year for Builder with no lab-tracking capability.
Cost: Basic $36/month (annual); no transaction fees.
For 30 students: $36/month flat. Cleaner UX than Teachable. Zero transaction fees.
Limitations: Same as Teachable — no lab tracking, no community, no external tool integration for completion proof.
Verdict: Better value than Teachable but same limitations. Still not built for hands-on lab programs.
Cost: Hobby $9/month (10% transaction fee); Pro $99/month (2.9% transaction fee).
What it does: Community (Facebook-group style) + courses + gamification (points/leaderboards) + events calendar. Alex Hormozi's preferred platform — designed for exactly this: paid cohort programs with community.
For Jumpstart: The leaderboard and points system would drive engagement among high schoolers. Events calendar for Zoom classes is built-in. Course modules can be structured as weekly unlocks.
Limitations: No external lab tracking, no screenshot upload for completion proof. The community is Skool-hosted (skool.com/cloud-8-academy), not your domain.
Verdict: Best off-the-shelf option IF you want community + courses in one place without building. At Pro ($99/month) the 2.9% transaction fee on top of Stripe is redundant if you handle payments through Stripe already. Use Skool for community only and TITAN for lab tracking — hybrid approach.
Cost: $41/month (Courses plan, annual).
What it does: Community + courses. More customizable than Skool, supports native video, spaces, sub-groups.
Verdict: More expensive than Skool, less engagement-optimized for younger users. Skip.
Cost: Free (Substack takes 10% of paid subscriptions).
What it does: Newsletter-as-course — weekly email drops with content, links, assignments. No tracking, no community, no labs.
Verdict: Useful as a supplemental announcement channel. Not an LMS. Could be the "weekly parent update" channel.
Cost: Free.
What it does: Real-time community, voice channels for live sessions (backup if Zoom fails), weekly announcement channels, role-based access (Student, Intern, Alumni).
Verdict: Essential for program community regardless of LMS choice. Set up #week-X-labs, #wins, #capstone, #parent-updates channels. Harnoor pins weekly content drops as announcements.
Cost: $0 (TITAN already exists).
What it does: Full tile-based experience per the spec in Section C. Progress tracking, lab screenshot uploads, quiz engine, PM view — all in-house. Deep Stripe integration (enrollment status synced). Deep branding control.
Verdict: This is the recommendation. TITAN already has the bridge infrastructure. The /academy page is a new route, not a new system. Build the tile spec in Section C iteratively — Week 1–4 content first, then expand.
TITAN /academy — content delivery, tile tracking, progress, quizzes
Discord — community, real-time support, weekly announcements
Third-party labs — Replit, AWS Educate, HF Spaces (embedded/handoff cards)
Stripe — payments (already configured)
Resend — email automation (drip sequences)
Total incremental cost: $0/month to launch. No LMS subscription needed for Cohort 1.
---
Every person in the Cloud 8 orbit has a designation. The /pm dashboard surfaces everyone in one view.
{
"people": [
{
"id": "milana-001",
"name": "Milana [Last]",
"designation": "Intern",
"status": "active",
"commission_pct": 15,
"current_tasks": [],
"last_update": "2026-04-25"
},
{
"id": "student-001",
"name": "Student Name",
"designation": "Student",
"tier": "accelerator",
"cohort": "1",
"week": 3,
"progress_pct": 45,
"last_active": "2026-04-26"
}
]
}
Panel 1 — Roster Table
Columns: Name | Designation | Status | Current Task | Due Date | Last Update | Health
Health colors: Green (on track, last active ≤ 3 days ago, ≥ 80% progress) | Yellow (3–7 days since last active or 50–79% progress) | Red (>7 days inactive or < 50% progress)
Panel 2 — Weekly Friday Update Aggregator
Every team member (Intern + Students) submits a 5-line Friday update via a form at academy.cloud8data.com/friday:
1. What I completed this week
2. What I'm stuck on
3. My plan for next week
4. Hours spent (approx)
5. One thing that was great / one thing that sucked
Submissions appear in Panel 2 grouped by person. Harnoor reviews Sunday. Outstanding submissions auto-color red.
Panel 3 — Open Tasks
Aggregated view of all assignment tiles with status "submitted" — waiting for Harnoor's review. Click any row to open the student's submission inline.
Panel 4 — Milana's Funnel Dashboard
Specific sub-panel for the Intern role:
One-Click Reminder
Each red-health person has a "Send Reminder" button. Clicking it triggers a Resend email using the student's stored email + a template: "Hey [name], your coach noticed you haven't checked in this week. Here's your Week [X] lab waiting for you: [link]. Reply if you're stuck."
---
File: F:/TITAN/state/people.json
{
"designations": {
"Intern": {
"description": "Paid staff on profit-share. Runs marketing, sales, parent relationships.",
"template_page": "/people/intern",
"permissions": ["view_roster", "view_student_progress", "submit_friday_update", "view_own_commission"],
"commission_pct": 15
},
"Student": {
"description": "Enrolled high schoolers. Paying tuition or scholarship.",
"template_page": "/people/student",
"permissions": ["view_own_tiles", "submit_assignments", "view_own_progress", "access_discord"],
"tiers": ["foundations", "accelerator", "college-ready"]
},
"Future-Hire": {
"description": "Candidates under evaluation for Cloud 8 engineering or intern roles.",
"template_page": "/people/future-hire",
"permissions": ["view_assigned_tasks"],
"pipeline_stage": ["applied", "screened", "trial-project", "offer", "hired", "declined"]
},
"India-Pool": {
"description": "10 available offshore candidates for engineering work.",
"template_page": "/people/india-pool",
"permissions": ["view_assigned_tasks", "submit_work"],
"availability": "on-demand"
}
}
}
Each person also has their own page at /people/[id] showing:
---
Platform: Replit (primary), JupyterLite (backup for no-signup exercises)
Theme: "From zero to your first working program"
| Week | Topic | Video Resource | Lab | Assignment |
|------|-------|---------------|-----|------------|
| 1 | What is AI? + Python Setup | Harnoor's recorded intro (15 min, YouTube unlisted) | Fork Replit Python Starter: https://replit.com/templates/python | 50-word reflection: "What problem would you solve with AI?" |
| 2 | Variables, Functions, Lists | Harnoor's Week 2 video + optional Codecademy "Learn Python 3" (free course, codecademy.com/learn/learn-python-3) | Replit: write a Python function that calculates your GPA | Quiz: Python basics (5 questions) |
| 3 | Pandas DataFrames | YouTube: "Pandas in 10 minutes" (pandas official) | Colab notebook (Harnoor-authored): https://colab.research.google.com/github/[cloud8]/blob/main/week3-pandas.ipynb | Analyze a CSV of your choice — 3 insights in writing |
| 4 | NumPy + your first GitHub push | Harnoor's Week 4 video | GitHub Education signup + push your Week 3 analysis to a public repo | Submit GitHub repo URL |
Week 4 milestone: Every student has a GitHub account, a public repo, and has run real Python on real data.
Platform: Google Colab (primary), Hugging Face (intro), Replit (ongoing)
Theme: "Understanding how AI thinks — and directing it"
| Week | Topic | Video Resource | Lab | Assignment |
|------|-------|---------------|-----|------------|
| 5 | What is an LLM? Tokens, embeddings, prompts | Harnoor's Week 5 video (20 min) | Colab: tokenize text with HuggingFace tokenizers (https://huggingface.co/docs/transformers/tokenizer_summary) | Write 5 prompts for 5 different tasks. Which gave best results and why? |
| 6 | Prompt engineering — the Balance AI+Brain framework | Harnoor's live class replay | Replit Groq Chatbot: fork https://replit.com/@replit-matt/Chat-with-Groq?v=1 | Build a chatbot that answers questions about a topic you care about |
| 7 | Scikit-learn: train your first ML model | Harnoor's Week 7 video | Colab: train a decision tree on Titanic dataset (Kaggle public dataset, no account needed: https://www.kaggle.com/c/titanic/data) | Submit: model accuracy + 3 observations about what mattered most |
| 8 | HuggingFace Spaces — ship your first demo | Harnoor's Week 8 video | Create HF account + deploy a Gradio text-classification demo on Spaces: https://huggingface.co/blog/gradio-spaces | Submit your live HF Spaces URL |
Week 8 milestone: Every student has a live, public AI demo on Hugging Face Spaces. This is the first shareable artifact.
Platform: AWS Educate (foundation), AWS Skill Builder + Cloud Quest (primary labs)
Theme: "The infrastructure behind every AI product you use"
| Week | Topic | Video Resource | Lab | Assignment |
|------|-------|---------------|-----|------------|
| 9 | What is the cloud? AWS 101 | Harnoor's Week 9 video + AWS Educate Cloud Literacy badge pathway | AWS Cloud Quest: Cloud Practitioner — Mission 1 (skillbuilder.aws/cloudquest) | Screenshot of completed Cloud Quest mission + 100-word reflection |
| 10 | S3 + IAM + EC2 basics | AWS Educate missions (AWS console hands-on via Educate credits) | AWS Cloud Quest: Missions 2–3 | How would you use S3 in your capstone project? (100 words) |
| 11 | AWS AI services: Rekognition, Polly, Lex | Harnoor's Week 11 video on AWS AI services | AWS Educate: AI services intro lab (included in Educate pathway) | Build a 1-slide diagram of how an AWS AI service works |
| 12 | Putting it together — cloud-connected apps | Harnoor's Week 12 architecture overview | Deploy a simple static site to S3 (Harnoor-authored walkthrough in a Colab notebook) | Submit: your S3 bucket URL |
Week 12 milestone: Every student has used live AWS console, completed Cloud Quest missions, and deployed a static artifact to S3.
Platform: GitHub Codespaces (dev environment), Hugging Face Spaces (deployment)
Theme: "You build the thing. We mentor. You ship."
| Week | Topic | Activity |
|------|-------|---------|
| 13 | Capstone kick-off: project selection + scope | 1:1 mentor call (Accelerator/College-Ready tiers); project proposal submitted as Assignment tile |
| 14 | Build week 1: data pipeline + model prototype | Codespaces dev environment; weekly check-in on Discord; Harnoor reviews GitHub commits |
| 15 | Build week 2: front-end + HF Spaces deployment | Student ships working demo to HF Spaces; Harnoor gives written feedback |
| 16 | Presentation + graduation | Student records 5-min Loom demo of their project; submits GitHub repo + HF Space + Loom link; receives certificate |
Week 16 deliverables per student:
1. Public GitHub repo with full project (clean README, instructions, architectural notes)
2. Live Hugging Face Spaces demo URL
3. Loom demo video (5 min)
4. LinkedIn post (Harnoor reviews draft for Accelerator/College-Ready tiers)
5. Certificate of Completion (Cloud 8 Data / HM Tech Solutions LLC)
---
Milana runs the human funnel. TITAN runs the tracking + logistics backbone.
cloud8data.com/jumpstartToday panel:
Pipeline panel:
Enrolled students panel:
intern_attribution: milana)---
| Day | Action | Owner |
|-----|--------|-------|
| Mon Apr 28 | Register academy.cloud8data.com subdomain, point to TITAN /academy route | Harnoor |
| Mon Apr 28 | Create Discord server: "Cloud 8 Academy" — channels: #announcements, #week-1, #wins, #q-and-a, #parent-lounge | Harnoor |
| Tue Apr 29 | Build /academy skeleton: 3 tiles (1 video, 1 lab, 1 assignment) for Week 1 | FORGE (TITAN) |
| Tue Apr 29 | Create Replit Team for Education — invite Milana as co-admin | Harnoor |
| Wed Apr 30 | Set up 3 Colab notebooks (Week 3, 5, 7) in GitHub repo cloud8data/jumpstart-labs | Harnoor |
| Thu May 1 | Set up Hugging Face organization cloud8-academy for student capstone deployments | Harnoor |
| Thu May 1 | Test full enrollment flow: Stripe checkout → webhook → TITAN people.json entry → /academy access | Harnoor + TITAN |
| Fri May 2 | First Friday Update submitted (Harnoor to self, as a dry run of the system) | Harnoor |
End of Week 1 success condition: A student with zero prior knowledge can sign up on Stripe, land on /academy, watch the Week 1 video, fork the Replit starter, and mark it done — all within 60 minutes.
| Day | Action | Owner |
|-----|--------|-------|
| Mon May 5 | Milana returns from exams — onboard to /pm, Discord, Calendly | Milana + Harnoor |
| Mon May 5 | Enroll 3 students manually (free or scholarship) for dry run — can be family/friends | Harnoor |
| Tue May 6 | Run Week 1 dry-run live Zoom class (60 min) — follow demo class agenda (Section J, Runbook) | Harnoor |
| Wed May 7 | Collect feedback from 3 dry-run students on tile UX, Replit setup experience | Harnoor |
| Thu May 8 | Fix top 3 friction points from dry-run feedback | FORGE |
| Fri May 9 | Milana begins outreach — target: 20 parent contacts in Week 2 | Milana |
End of Week 2 success condition: 3 students have completed Week 1 tiles. At least 1 has submitted the Week 1 assignment. Milana has 20 contacts in her outreach pipeline.
| Day | Action | Owner |
|-----|--------|-------|
| Mon May 12 | Run first paid intro class (Zoom, 75 min) using live-demo-script.md | Harnoor |
| Mon May 12 | Activate EARLYBIRD50 coupon on Stripe — 24-hour window | Harnoor |
| Tue May 13 | Follow-up DM sequence via Milana to all attendees | Milana |
| Wed May 14 | Target: 7 enrollments by end of day | Harnoor + Milana |
| Thu May 15 | Onboard 7 students: Discord roles, /academy access, Replit Team invite | Milana |
| Fri May 16 | First full Friday Update from all 7 students | All |
End of Week 3 success condition: 7 paid students enrolled. At least 4 have completed Week 1 tiles. EARLYBIRD50 has driven 4+ redemptions.
| Day | Action | Owner |
|-----|--------|-------|
| Mon May 19 | Second paid intro class (for late arrivals) — targeting seats 8–20 | Harnoor |
| Tue May 20 | Official Cohort 1 Week 1 live class (all enrolled students, Zoom) | Harnoor |
| Wed May 21 | Target: 15–20 enrolled students | Harnoor + Milana |
| Thu May 22 | Collect parent emails from Milana's pipeline; send parent welcome sequence | TITAN Resend |
| Fri May 23 | Weekly Friday Update aggregated in /pm — first formal cohort check-in | All |
End of Week 4 success condition: 15–20 paid students in Cohort 1. First live class completed. Parent emails collected. Revenue: 15 × $297 (Foundations avg) = $4,455 MRR minimum.
---
Why this lab:
1. Zero installation — runs in any browser, any device
2. Free — Replit free account + Groq free API key (no credit card)
3. Result is a working AI chatbot in under 20 minutes — maximum "wow" factor for parents watching
4. Connects directly to the Balance AI+Brain framework Harnoor demonstrates in the live-demo-script
5. The code is real Python — students feel like real engineers, not like they clicked through a tutorial
Template: https://replit.com/@replit-matt/Chat-with-Groq?v=1#main.py
Groq free API: https://console.groq.com (free account, no credit card, instant API key)
0:00 – 0:05 Welcome + who's in the room (name + grade poll in chat)
0:05 – 0:15 Harnoor context (2-min version of live-demo-script Problem Block)
0:15 – 0:20 Explain what we're building: "In the next 35 minutes, you're going to
build and run a real AI chatbot. Not watch someone else do it. You."
0:20 – 0:25 Setup:
Step 1: Open replit.com, create free account (2 min)
Step 2: Go to console.groq.com, create free account, copy API key (3 min)
0:25 – 0:40 Fork + run:
Step 3: Fork https://replit.com/@replit-matt/Chat-with-Groq?v=1
Step 4: Add GROQ_API_KEY to Replit Secrets
Step 5: Hit Run
Step 6: Type your first message to the chatbot
[Harnoor live-codes this while students follow]
0:40 – 0:50 Customize it: change the system prompt to make the chatbot answer
only about a topic the student cares about (sports, music, gaming)
"See? You just directed an AI. That's the skill."
0:50 – 0:55 Show the code: "Here's where the chatbot's 'personality' lives.
Here's where your message goes in. Here's where the response comes back.
This is the architecture of every AI product you use."
0:55 – 1:00 CTA: "In 16 weeks, you'll build something 10x more sophisticated.
You just wrote your first AI app. Enrollment link is in the chat."
Instructor prep:
Backup if Groq signup is slow: Pre-build a hosted version on Replit with a read-only link that students can interact with (no signup needed) — students type messages and see responses without needing their own API key. Then frame personal setup as "your homework after class."
---
Framework last updated 2026-04-27. Owner: Harnoor Minhas. Builds on: live-demo-script.md, parent-packet.md, stripe-products.md.