For working engineers · September 2026 cohort

Applied AI Engineering: the deep dive

Evolve into a modern AI-native builder shipping production-grade AI systems.

Key details

Duration
Six weeks, 12 sessions, ~5 hrs/week
Cohort
15 seats — capped for direct instructor access
Format
Live weekend classes: 50% theory, 50% build, plus a weekly assignment
Prereq
Write Python or JS/TS; no AI/ML background needed
Recordings
All sessions recorded for enrolled students
Instructor access
Office hours + live 1:1 review of your shipped feature

6/15 seats filled · applications close Aug 20, 2026.

Curriculum

One codebase, six weeks — it grows every class. You extend the same fork from week 1 through week 6.

Week 1 · Orientation & first extension

Theory — intro to applied generative AI

Hands-on — fork the starter, run it locally, ship a first real extension — and get familiar with AI-augmented software development

Week 2 · Structured outputs & tool calling

Theory — structured outputs, tool-calling patterns, debugging prompts like code

Hands-on — add a tool-calling feature against the existing data model

Week 3 · Retrieval & grounding

Theory — chunking, embeddings, retrieval quality, when RAG is the wrong tool

Hands-on — ground the feature in the app's own data

Week 4 · The analyst agent

Theory — agent orchestration, multi-step planning, safe failure modes

Hands-on — ship an analyst agent inside the product

Week 5 · Evaluation & observability

Theory — monitoring, observability, fallbacks, and evaluation

Hands-on — eval harness, golden datasets, fallback, tracing

Week 6 · Hardening & ship review

Theory — overview of access control, data handling, safe rollout, and cost/latency tradeoffs

Hands-on — harden your branch and walk through it live for your portfolio

Don't feel ready for the advanced course? Check out my beginner friendly course for building the basics.

Who this is for

Good fit

  • You write Python or JS/TS day to day.
  • You've used LLM APIs but never shipped an agentic system.
  • You want to move into AI-focused engineering roles.
  • You'd rather build one real project than watch slides.

Not a fit

  • You're brand new to programming.
  • You want a "prompt tips" course.
  • You want a certificate over the skill.
  • You can't spare ~5 hrs/week for 6 weeks.

Pricing

Online cohort · September 2026

₹35,000 ₹60,000

The online price going forward — the ₹60,000 in-person format is being retired. Up to 90% refund any time before your 3rd class, no questions asked.

You'll get

  • 40+ hours of live weekend classes taught by me
  • Lifetime access to recordings
  • The starter repo and your extended fork, yours to keep
  • Office hours + live 1:1 review of your shipped feature
  • The cohort network

Most employers cover this as learning & development — you'll get a receipt for reimbursement.

Who's teaching

Rahul Gupta, PhD

I've founded and led the AI team at two US startups — TeamOhana in workforce management and Ariksa in cloud security — building the production AI systems those companies sell. This cohort teaches the kind of engineering behind that work, grounded in a PhD from IISc Bangalore.

  • PhD, Computer Science — IISc Bangalore
  • Published at AAAI & NeurIPS · 1000+ citations
  • Founded & led AI at two US startups — TeamOhana, Ariksa
  • 5+ revenue-generating AI systems in production in the last 2 years
  • 500+ engineers trained
  • Full background: Home

FAQ

What if I miss a live session?
Every session is recorded and shared with the cohort. Live sessions are where most of the value is, though — plan around them if you can.
Do I need prior AI/ML experience?
No. Be a working programmer; everything about LLMs and agents is taught from first principles.
What happens after I apply?
A short fit call before you pay anything. Applications close Aug 20, 2026, or once all 15 seats are filled — whichever comes first.
Why only 15 seats?
Office hours and a live 1:1 review only work at a small cohort size. 15 is the cap, not a soft target.
What's the refund policy?
Up to 90% refund any time before your 3rd class. After that, seats can't be backfilled.
What will I actually have at the end?
A live fork of a real product with an analyst agent feature you built and shipped yourself — plus direct feedback on it.
Do I need to know the original production system?
No. Week 1 walks you through the architecture and data model before you touch any code.
Can I get this reimbursed by my employer?
Yes — most people expense this as L&D. Payment via UPI or bank transfer, with a receipt for your reimbursement request.

Apply

15 seats. September 2026. Applications close Aug 20.

Just a few details to help me decide the fit. No payment required until acceptance.

6/15 seats filled · applications close Aug 20, 2026.