Cohort 07 starts November 4 · 32 seats only

Go from Full Stack to AI Engineer in 6 weeks

A live, code-first cohort that takes working full stack developers from "I've called an OpenAI API once" to shipping production LLM systems — RAG pipelines, evals, agents, and a portfolio hiring managers actually open.

  • 6-Week Live Build Track — Every Saturday we ship a real feature together. By week 12 you have four production-grade AI apps in your GitHub.

  • Code Reviews From a Staff AI Engineer — Submit PRs each week. Get line-by-line feedback on prompts, retrieval, latency, and architecture decisions.

  • Hiring-Manager Portfolio Kit — Resume rewrite, LinkedIn positioning, two mock interviews, and direct intros to our 40+ hiring partners.

847 developers enrolled across 6 cohorts
4.9/5 from 312 reviews

Get the Free AI Engineer Starter Kit

The 40-page playbook + 5 starter repos we use in week 1. Join 12,400+ devs on the list.

2,847 downloads this month

No spam. Unsubscribe anytime. We also send one weekly AI engineering digest.

Our graduates ship AI at

Stripe Notion Vercel Linear Anthropic Ramp Scale

Cohort outcomes · 2025-2026

Numbers that matter to your career

We track every graduate for 12 months after the cohort ends. Here's the median, not the marketing version.

$42K

Median salary increase

↑ 38% YoY

11 wks

Median time to first AI role

↓ 7 wks vs 2023

94%

Land an AI role within 6 months

↑ Stable

4.9/5

Average NPS across 6 cohorts

✓ Verified

Career Note — The biggest salary jumps went to grads who completed all four portfolio projects (not just watched the videos). The cohort is structured so finishing is the default, not the exception.

What grads say

They came in shipping REST APIs. They left building AI systems.

"I'd been a senior full stack dev for six years and felt invisible to every AI job posting. Eleven weeks after the cohort I had three offers — including the staff AI role at a YC company. The portfolio projects were the differentiator."

Marcus Chen

Staff AI Engineer · YC Series B startup

"The RAG phase alone was worth ten times the tuition. I'd watched four YouTube series and read every blog post — none of them showed me how to actually evaluate retrieval quality. The eval harness we built in week 7 is now running in our CI at work."

Priya Ravichandran

Senior Engineer · Fortune 500 retailer

"I'm a frontend-leaning dev who'd never touched Python. Twelve weeks later I shipped a multi-agent system that's now powering our customer onboarding. Arjun's code reviews are unreasonably specific and that's exactly what you need."

Daniel Kim

AI Engineer · healthcare SaaS

The 6-week build track

What you'll ship, week by week

No toy demos. Every module ends with a deployed, tested project that belongs on your résumé.

Phase 1 · Weeks 1–2

Foundations

The mental models every AI engineer uses daily.

  • ·Tokens, context windows, and embeddings
  • ·Prompt engineering as a first-class skill
  • ·Structured outputs, tool use, function calling
  • ·Ship: A production-grade chat API
Most grads cite this
Phase 2 · Weeks 3–4

RAG & Retrieval

The skills that show up in 80% of AI engineering job posts.

  • ·Vector databases and chunking strategies
  • ·Hybrid search, reranking, query rewriting
  • ·Evaluation harnesses and regression testing
  • ·Ship: A multi-tenant RAG support bot
Phase 3 · Weeks 5-6

Agents & Production

Where interviews separate juniors from staff-level.

  • ·Agent loops, planning, and tool orchestration
  • ·Latency, caching, and cost optimization
  • ·Observability, guardrails, evals in CI
  • ·Ship: An autonomous research agent
Download the full syllabus

8 yrs

building production LLM systems at scale

Your instructor

Rajani Meka

Staff AI Engineer · previously led the LLM platform team at a Series C fintech and in Law firm

I built the review queues your bank's app uses to spot fraud, and the internal copilot our 3,000 engineers shipped 11M prompts through last quarter. I've hired eleven AI engineers in the last two years — and rejected about four hundred. I built this course to be the exact curriculum I wish those four hundred had.

Former

ML Lead, Plaid

Speaker

PyCon · ODSC · AI Engineer Summit

Built

Llmkit (4.1k★ OSS)

"The gap between 'I can call an LLM' and 'I can ship one to production' is enormous. Most courses stop at the first half. We don't."

How the cohort runs

Six hours a week. That's the whole ask.

Designed for full stack developers with a day job. No fluff, no 4-hour lectures, no "figure it out yourself" homework.

01

Tue, Th. Live Lecture

90 minutes. New concept, code walkthrough, Q&A. Replay posted by midnight.

02

Thu · Office Hours

60 minutes of small-group debugging. Bring your actual code, leave with a PR merged.

03

Recorded Vidoes

Learn with recoded play and build the same feature in real time. You ship something to prod by Sunday. 

04

Async · Code Review

Submit your PR by Sunday. Get line-by-line feedback from Rajani by Wednesday.

Common questions

The stuff you actually want to know

Still have a question? Email supprt@edutva.com — Rajani answers personally within a business day.

I have a full time job. Realistically, how much time will this take?

About six hours a week. One 90-minute live lecture (Tuesday), one 60-minute office hours (Thursday), and roughly 2.5 hours of async work on the weekend project. We built the schedule around people with day jobs because most of our students have one.

Do I need to know Python deeply, or is JS / TS experience enough?

Roughly a third of our grads are TypeScript-first. We include a 2-week Python primer in week 1 — enough to be productive with FastAPI, async, and Pydantic. After that we focus on the AI engineering patterns, not the language. If you can ship a production API in Node, you can ship one in Python by week 3.

What happens after the 6 weeks end?

You keep lifetime access to the recordings, starter repos, and the private alumni community. The hiring partner intros run after graduation. About a third of each cohort stays active in the alumni Slack long-term — it's become the most useful AI engineering room on the internet.

Is there a job guarantee?

No fake guarantees. What we offer: 94% of grads who complete the full program land an AI role within 6 months. We publish a 30-day-out outcome report every quarter. The best predictor of outcome isn't the cohort — it's whether you finish the projects.

What if I can't make the live sessions?

Recordings go up within four hours of every session. Office hours have two time slots (one APAC-friendly, one Americas-friendly). About 40% of students attend async-only and still finish strong — the projects are the substance, the live calls are the bonus.

How is this different from the free content on YouTube?

Two things. First, accountability: a fixed cohort, weekly PRs, and a real human reviewing your code. Second, depth: we go past the "wrap an OpenAI call" tutorials into the production patterns — evals, observability, cost optimization, agent architectures — that almost no free content covers well.

Cohort 07 · Starts November 4

6 weeks from now, you could be shipping AI systems — not watching tutorials about them.

Limited seats. Doors close November 1, or when the cohort fills — whichever comes first.

  • Live Classes and Recorded Session will be proved
  • Partly payment plans available at checkout
  • Group rates for teams of 3+ — email us about it

Get the full syllabus + pricing

No card needed

We'll email you the syllabus, the four portfolio project specs, and a 5-minute video walkthrough of how the cohort actually runs.

847 developers enrolled across 6 cohorts