What's Inside the Learning Portal
4 courses, 18 lessons, 5 modules, interactive MDX components, progress tracking, and two enrollment flows — here's how it all works.
What's Inside the Learning Portal
The learning portal is the heart of this platform. Here's everything it offers — the courses, the content system, the enrollment flows, and how progress tracking works.
The Courses
Deep learning, transformers, LLMs, and prompt engineering from scratch. The flagship course — structured as a proper 3-hour curriculum.
Kubernetes, CI/CD pipelines, Docker, Terraform, and production-grade infrastructure.
End-to-end ML pipelines, model serving, experiment tracking, and monitoring in production.
System design, DSA, behavioural rounds, and mock-interview strategies. The only paid course.
AI Bootcamp — Deep Dive
The AI Bootcamp is the most comprehensive course on the platform. It's organized into 5 modules, each covering a distinct area.
Course overview, objectives, and learning roadmap
Foundations of generative models — GANs, VAEs, diffusion models
7 lessons — from "What is an LLM?" to neural networks, transformers, and model architecture
Zero-shot, few-shot, chain-of-thought, and advanced prompting techniques
Hands-on projects applying AI to emails, resumes, interviews, and daily productivity
Interactive Content
Plain Markdown isn't enough for a learning platform. Every lesson uses custom React components embedded in MDX for rich, interactive content.
Info, warning, tip, and fun-fact boxes for highlighting important concepts
Hands-on practice blocks with clear instructions
Side-by-side before/after concept comparisons
Structured multi-row comparison tables
Key points to remember after each section
Thought-provoking questions to deepen understanding
Visual data grids — like the ones you're reading right now
Every piece of structured data on this platform uses InfoGrid/InfoCard instead of Markdown tables. They look better, work on mobile, and support icons and subtitles.
Content Structure
Lessons are .mdx files stored in the repo. The folder structure maps directly to URLs.
content/learning/
├── devops/
│ ├── k8s-lesson-1.mdx → /learning/devops/k8s-lesson-1
│ └── k8s-lesson-2.mdx → /learning/devops/k8s-lesson-2
├── ai-bootcamp/
│ ├── about-course/
│ │ └── 01-welcome.mdx → /learning/ai-bootcamp/about-course/welcome
│ ├── generative-ai/
│ │ └── 01-what-is-gen-ai.mdx → /learning/ai-bootcamp/generative-ai/what-is-gen-ai
│ ├── large-language-models/
│ │ ├── 01-what-is-llm.mdx → /learning/ai-bootcamp/large-language-models/what-is-llm
│ │ └── ...7 lessons total
│ ├── prompt-engineering/
│ │ └── ...
│ └── ai-in-action/
│ └── ...
├── mlops/
│ └── mlops-intro.mdx → /learning/mlops/mlops-intro
└── interview-prep/
└── interview-intro.mdx → /learning/interview-prep/interview-intro
Files use prefixes like 01-, 02- to control lesson order. A stripOrderPrefix() utility removes the number for clean URL slugs — so 03-neural-networks.mdx becomes /learning/.../neural-networks.
Progress Tracking
Every lesson has a "Mark Complete" button at the bottom. Progress is stored in localStorage — no backend calls, no latency.
Shows overall course completion percentage at the top of the course page
Visual list of all lessons with completed/in-progress indicators
Per-module lesson counts and completion status for module-based courses
Mark a lesson complete → progress updates immediately, no page reload needed
Enrollment Flows
The course catalog is open to all signed-in users — everyone can browse courses and see what's available. Access is granted through two flows depending on whether the course is free or paid.
Free Courses
Student clicks "🔓 Request Access" on the course page
I get a Gmail notification with the student's details and a one-click approve button
I tap the approve button from my phone — HMAC-signed, no login required
Student refreshes the page and they're in — full access to all lessons
Paid Course (₹125)
Student scans the UPI QR code (amount pre-filled) or pays manually using the UPI ID
Student enters their UPI transaction ID in the payment form
I get an email with the transaction ID — I cross-check in my UPI app and tap approve
Student is enrolled instantly — full course access on refresh
Whether free or paid, the final step is always a one-click HMAC-signed approve link in my email. The student's enrollment is stored in Clerk's publicMetadata — no database writes, no sync issues.
Curated Programs
Not everyone needs the full course. Sometimes a learner needs just Kubernetes + CI/CD from DevOps and a few AI lessons — no Ansible, no advanced prompting.
The curated programs system lets me assemble a hand-picked lesson playlist from any course, scoped to a specific person or cohort.
Each program is a named subset of existing lessons — no content duplication, just references to lesson paths
Progress tracking is per-program — if a program has 9 lessons, the progress bar shows 0/9
Programs live behind a single "Curated Programs" card on the catalog — click it to see all available programs at /programs
Visual wizard at /admin/programs — browse lessons, pick the ones you want, download a JSON config file. No code editing needed.
Connect With Me
Have questions about a course or want to discuss a topic in depth? You can schedule a 15-minute call — powered by Cal.com with automatic Google Meet links.
What's Coming
Auto-generated completion certificates when all lessons are marked done
Per-lesson comment threads for questions and community interaction
Optimized reading layout for phone screens
Expanding the catalog with new topics and deeper content