Program map

18weeks5phases

0P1–P2Pre-courseZero to coder
11–6Full-stack MERNBasic
27–11AI engineering in TypeScriptIntermediate
312–14Agents and productionAdvanced
415–16Mastery and launchMaster

Waitlist openLive program

MERN Full Stack AI Engineer

In 18 weeks — a 2-week pre-course plus a 16-week core — a student goes from zero to building and shipping full-stack AI products on the MERN stack, entirely in JavaScript/TypeScript, and leaves with either a launched product or a job-ready portfolio.

Who it's for: beginners who want one language for everything, and web developers moving into AI. Students who already code skip the pre-course after an entry test.

Stack: MongoDB, Express, React, Node.js — plus TypeScript and Next.js, since most modern MERN teams now ship React through Next.js. The AI layer is built in TypeScript too: Vercel AI SDK, OpenAI/Anthropic/Gemini SDKs, LangChain.js, LangGraph.js, Mastra, MongoDB Atlas Vector Search.

18 weeksincl. 2-week pre-course
5 phasesZero to coder → Master
Mon–ThuLive class (2–2.5 hrs)
FriReading day
SatAssignment day

Internship included

We provide an internship as part of the program. You work directly on our products, alongside the team that builds them.

Outcomes by Week 16

Builds and deploys full-stack MERN apps with auth, payments, background jobs, and tests
Adds AI to any product: streaming chat, structured output, tool use, RAG, agents, MCP, voice and vision
Measures AI quality with evals and runs it safely in production
Uses AI coding tools (Claude Code, Cursor, Codex, Copilot) to ship fast without losing understanding
Has 4 portfolio projects plus 1 capstone live on public URLs
Founder track: a launched MVP with real users
Career track: profiles ready, applications out, mock interviews done

₹49waitlist

See curriculum

Curriculum · 18 weeks · 5 phases

Week by week, day by day

Weekly rhythm

Mon–ThuLive class (2–2.5 hrs): You teach the week's 4 topics: concept, live build, students build the next step themselves
FriReading day: Core + 2–3 latest articles, a 200-word summary, and one small experiment on their own project
SatAssignment day: The week's build + a short no-AI debugging task; GitHub link + demo video
SunOff: Rest; optional doubt-clearing call

AI usage rules by phase

P1–P2, Weeks 1–2: No AI code generation. AI only for concepts and explaining errors, with learning mode on
Weeks 3–11: Generate-then-explain: every AI-written block is explained in the PR or demo video
Weeks 12–16: Full agentic tools, plus one no-AI debugging drill each week

How students learn here

Monday recall quiz (10 min): on last week and 3 weeks back
I do / We do / You do: in every live class

Cohort setup

Pods of 4–5 with a buddy; mentor groups of 25 or fewer
Unguided builds: Project 3 and the capstone get a goal and rubric, not step-by-step instructions

Phase 0Zero to coder

Pre-course

Weeks P1–P2

By the end of Week P2, a complete beginner can build and publish an interactive web page with HTML, CSS, and JavaScript, using Git and the terminal.

Day 1 · Monday · Live class (2–2.5 hrs)
How software works and setting up
CPU, RAM, storage, OS, files, processes
VS Code, Node.js (via nvm), Git, a browser with DevTools; WSL for Windows
pwd, ls, cd, mkdir, touch, cp, mv, rm; paths; environment variables and PATH
asking clear questions, pasting full errors, verifying answers
Day 1 · Monday · Live class (2–2.5 hrs)
JavaScript basics I
let, const, numbers, strings, booleans, null, undefined; operators; type coercion
methods, template literals
if/else, switch, for, while, for...of
the browser console and Node error messages

Phase 1Basic

Full-stack MERN

Weeks 1–6

By Week 6, every student ships a deployed full-stack MERN app with auth, real-time updates, and tests — built before any AI is added, so they understand every layer they'll later automate.

Day 1 · Monday · Live class (2–2.5 hrs)
How JavaScript really works
var vs let/const, temporal dead zone
what they are, practical uses (counters, memoisation, module pattern)
rules, call/apply/bind, arrow functions and this
prototype chain, class syntax, inheritance, getters/setters, private fields
Day 1 · Monday · Live class (2–2.5 hrs)
TypeScript fundamentals
catching bugs before runtime, editor support
primitives, arrays, tuples, objects, type vs interface, optional and readonly
typeof, in, literal types
generic functions and types
tsconfig, strict mode, tsx/ts-node
Day 1 · Monday · Live class (2–2.5 hrs)
React fundamentals
components, JSX, one-way data flow
lifting state up, controlled inputs
conditional rendering, lists and keys
Vite + React + TypeScript
Day 1 · Monday · Live class (2–2.5 hrs)
Node.js internals
libuv, non-blocking I/O, worker threads
reading large files, piping
building a server with the http module
Day 1 · Monday · Live class (2–2.5 hrs)
MongoDB fundamentals
MongoDB vs SQL, when to use which
Atlas, Compass, mongosh
insert, find, update, delete; query and update operators
Day 1 · Monday · Live class (2–2.5 hrs)
Connecting frontend and backend
login, refresh tokens, protected routes
02Errors and loading states across the stack
pnpm workspaces or Turborepo, shared types

Phase 2Intermediate

AI engineering in TypeScript

Weeks 7–11

By Week 11, students add streaming AI, tools, and RAG on MongoDB Atlas to their own MERN apps, measure quality with evals, and turn it into a paid AI SaaS.

Day 1 · Monday · Live class (2–2.5 hrs)
Inside an LLM (no heavy maths)
tokenization, why tokens are not words, token cost of Indian languages
the intuition behind transformers
pre-training, instruction tuning, RLHF/DPO, reasoning models
next-token prediction, temperature and top-p, context windows, KV cache
hallucination, knowledge cutoff, sycophancy
Day 1 · Monday · Live class (2–2.5 hrs)
Prompt engineering
system prompts, clear instructions, few-shot examples, XML tags
chain-of-thought, extended thinking, prompt chaining, self-critique
templates, versioning, storing prompts outside code
promptfoo
Day 1 · Monday · Live class (2–2.5 hrs)
Embeddings
meaning as vectors, cosine similarity
OpenAI, Voyage AI, Cohere; local embeddings with Transformers.js
embed and embedMany in the AI SDK, batching, cost
semantic search, recommendations, deduplication
Day 1 · Monday · Live class (2–2.5 hrs)
Evals foundations
vibes vs measurement
reading traces, grouping failures into categories
golden sets, edge cases, synthetic data generation
Day 1 · Monday · Live class (2–2.5 hrs)
Generative UI
rendering tool results as React components
attachments, suggestions, editing AI output, feedback buttons
images and files in chat

Phase 3Advanced

Agents and production

Weeks 12–14

By Week 14, students build agents in TypeScript with LangGraph.js and Mastra, ship an MCP server, add voice or vision, and run everything safely at scale — then kick off their capstone.

Day 1 · Monday · Live class (2–2.5 hrs)
Agent foundations
an LLM using tools in a loop toward a goal
chaining, routing, parallelisation, orchestrator-workers, evaluator-optimiser
ReAct, plan-and-execute, reflection
04When NOT to build an agent
Day 1 · Monday · Live class (2–2.5 hrs)
Model Context Protocol (MCP)
one standard way to plug tools into any AI app
hosts, clients, servers; tools, resources, prompts; stdio vs Streamable HTTP
03Build a server with the TypeScript SDK, test with MCP Inspector
deploying, OAuth 2.1 authorisation
tool poisoning, prompt injection through tool results, least privilege
Day 1 · Monday · Live class (2–2.5 hrs)
Observability and security
Langfuse, OpenTelemetry; time to first token, p95 latency, cost, error rates
sampling live traffic, feedback loops, alerts
direct and indirect prompt injection, jailbreaks, data leaks; OWASP Top 10 for LLM Apps
moderation APIs, Llama Guard, output validators, PII redaction
promptfoo red-team mode
India's DPDP Act, EU AI Act overview

Phase 4Master

Mastery and launch

Weeks 15–16

Students build their capstone while learning both how to launch a product and how to get hired, then present at demo day.

Day 1 · Monday · Live class (2–2.5 hrs)
Founder track: positioning and launch
one-line pitch, ideal customer, competitor map
headline, demo video, social proof, call to action
usage-based vs seats, free tiers, margins after model costs; India vs global pricing
Product Hunt, Peerlist, Show HN, Reddit, niche communities
X, LinkedIn, short demo videos
Day 1 · Monday · Live class (2–2.5 hrs)
Capstone polish
01Evals, tracing, and a security pass
02README, architecture diagram, and 3-minute demo video

Capstone requirements

A MERN (or Next.js + MongoDB) app solving a real problem for a named user group
Uses at least 2 of: RAG on Atlas Vector Search, tool use, agents (LangGraph.js or Mastra), MCP, voice or vision
Has evals, tracing, guardrails, and a cost estimate per user
Auth, payments in test mode, deployed on a public URL
README, 3-minute demo video, and a published case study

Portfolio at graduation

01Full-stack MERN app (no AI)React, Express, MongoDB, auth, real-time, deploymentWeek 6
02AI expense trackerVision, structured output, tool callingWeek 8
03RAG product → AI SaaS (unguided)Atlas Vector Search, evals, billingWeek 10–11
04MCP server + agent, or voice agentAgents, MCP or realtime voiceWeek 13
05CapstoneEverything, production-readyWeek 16

Grading: 60% weekly assignments, 30% capstone, 10% reading summaries and participation.

Complete MERN Full Stack AI Engineer skills map16 areas

Web basicsWeek P1–P2: Terminal, Git/GitHub, HTML, CSS (Flexbox, Grid), DOM, HTTP, JSON, fetch
JavaScriptWeek 1: Closures, this, prototypes, event loop, promises, async/await, ES modules, npm/pnpm, Vite, ESLint, Vitest
TypeScript and workflowWeek 2: Generics, narrowing, utility types, Zod, GitHub Actions, PR reviews, AI coding tools intro
ReactWeek 3: Hooks, React 19 actions, React Router, React Hook Form, TanStack Query, Zustand, Tailwind, shadcn/ui, Testing Library
Node and ExpressWeek 4: Event loop, streams, middleware, REST design, Swagger, Multer, JWT, OAuth, RBAC, helmet, rate limiting, Supertest
MongoDBWeek 5: CRUD, Mongoose, schema patterns, aggregation, indexes, transactions, change streams, Atlas Search, Redis
Full-stack and DevOpsWeek 6: Next.js App Router, server actions, Socket.io, SSE, BullMQ, Docker, Vercel, Render, AWS EC2, Nginx, PM2, Sentry
LLM foundations and APIsWeek 7: Tokens, attention, training, sampling; OpenAI/Anthropic/Gemini SDKs, Vercel AI SDK, OpenRouter, Ollama
Prompts, structure, toolsWeek 8: Prompting, context engineering, prompt caching, generateObject, Zod schemas, tool loops, LangChain.js
RAG on MongoDBWeek 9: Embeddings, Voyage/OpenAI/Cohere, Atlas Vector Search, chunking, Firecrawl, hybrid search, reranking, citations
Evals and observabilityWeek 10: Error analysis, LLM-as-judge, promptfoo, Evalite, Braintrust, Langfuse, LangSmith, semantic caching
AI SaaSWeek 11: Generative UI, Clerk/Auth.js/Better Auth, multi-tenancy, usage metering, Razorpay, Stripe
AgentsWeek 12: ReAct, agent loop from scratch, LangGraph.js (checkpointers, interrupts), Mastra, OpenAI Agents SDK, Claude Agent SDK
MCP, multi-agent, multimodalWeek 13: MCP TypeScript SDK, OAuth, A2A, Playwright, Stagehand, E2B, Inngest, vision, image generation, Realtime API, LiveKit, Vapi
ProductionWeek 14: Prompt injection, OWASP LLM Top 10, guardrails, scaling Node and MongoDB, k6, CI eval gates, fine-tuning overview
Product and careerWeek 14–15: Validation, MVP scoping, AI system design, launch, SEO, resume, interviews, job platforms

MERN Full Stack AI Engineer

Who it's for: beginners who want one language for everything, and web developers moving into AI. Students who already code skip the pre-course after an entry test.

18 weeksincl. 2-week pre-course

5 phasesZero to coder → Master

Mon–ThuLive class (2–2.5 hrs)

FriReading day

SatAssignment day

Internship included

We provide an internship as part of the program. You work directly on our products, alongside the team that builds them.

Outcomes by Week 16

Builds and deploys full-stack MERN apps with auth, payments, background jobs, and tests
Adds AI to any product: streaming chat, structured output, tool use, RAG, agents, MCP, voice and vision
Measures AI quality with evals and runs it safely in production
Uses AI coding tools (Claude Code, Cursor, Codex, Copilot) to ship fast without losing understanding
Has 4 portfolio projects plus 1 capstone live on public URLs
Founder track: a launched MVP with real users
Career track: profiles ready, applications out, mock interviews done

MERN Full Stack AI Engineer

Week by week, day by day

Weekly rhythm

AI usage rules by phase

How students learn here

Cohort setup

Pre-course

How software works and setting up

JavaScript basics I

Full-stack MERN

How JavaScript really works

TypeScript fundamentals

React fundamentals

Node.js internals

MongoDB fundamentals

Connecting frontend and backend

AI engineering in TypeScript

Inside an LLM (no heavy maths)

Prompt engineering

Embeddings

Evals foundations

Generative UI

Agents and production

Agent foundations

Model Context Protocol (MCP)

Observability and security

Mastery and launch

Founder track: positioning and launch

Capstone polish

Capstone requirements

Portfolio at graduation

MERN Full Stack AI Engineer

Week by week, day by day

Weekly rhythm

AI usage rules by phase

How students learn here

Cohort setup

Pre-course

How software works and setting up

JavaScript basics I

Full-stack MERN

How JavaScript really works

TypeScript fundamentals

React fundamentals

Node.js internals

MongoDB fundamentals

Connecting frontend and backend

AI engineering in TypeScript

Inside an LLM (no heavy maths)

Prompt engineering

Embeddings

Evals foundations

Generative UI

Agents and production

Agent foundations

Model Context Protocol (MCP)

Observability and security

Mastery and launch

Founder track: positioning and launch

Capstone polish

Capstone requirements

Portfolio at graduation