Working IT professionals switching into AI careers
Applied Generative AI
For developers moving into AI roles: build and ship LLM, RAG and agent systems in one month.
Monday to Friday evenings, scheduled around working hours
Four full-day Saturday labs, ending in a Demo Day — in the Himayat Nagar classroom for the Hyderabad batch, on Google Meet for the online batch. Included in the 24 sessions, not additional to them.
You should already be comfortable writing code.
What you will build
3 things that exist when the batch ends
Not slides, not a certificate on its own. Work that can be opened, run and interrogated by someone deciding whether to hire you.
A retrieval system that returns the right passage
Retrieval over a real document set, built end to end — including the chunking and ranking work that the tutorials skip and that decides whether any of it works.
- What it proves
- You can make retrieval hold up against documents that are long, messy and inconsistent, which is all documents.
- You leave with
- The service, its repository, and an eval run showing what measurably improved when you fixed the chunking.
- Python
- An LLM API
- A vector store
- An eval harness
An agent that finishes a multi-step task
Function calling, tool use and guardrails — an agent that reads, decides, acts, and stops when it is supposed to.
- What it proves
- You can build an agent that is bounded and observable, instead of one that demos well and loops in production.
- You leave with
- A running agent, the traces it emits, and the limits that hold it in.
- Function calling
- Tool definitions
- Guardrails
- Tracing
The deployed capstone
Built across the four Saturday labs: your own AI application, deployed, monitored, with its cost per request known rather than estimated.
- What it proves
- You can answer the three questions a team asks before anything ships — what does it cost, how slow is it, and how will we know when it breaks.
- You leave with
- A live application, a dashboard, and a cost-per-request figure you can defend to a finance team.
- Deployment
- Monitoring
- Cost instrumentation
Project briefs are finalised with your trainer at the start of each batch. The capability each one builds is fixed; the specific brief can flex to the cohort.
What you will be able to do
Outcomes
- Work fluently with an AI coding assistant inside a real codebase
- Design and build a retrieval-augmented system against your own documents
- Build an agent that calls tools and completes a multi-step task
- Evaluate an AI feature properly instead of judging it by vibes
- Reason about cost, latency and failure modes before shipping
- Leave with a deployed project and the repository behind it
The problem
You can already code. That is exactly why this is frustrating
You have read about RAG and agents, maybe wired up an API call, and you can tell the difference between a demo and a system. What you have not had is a month of supervised time to build the real thing — with the chunking that actually works, the evaluation that catches regressions, and the cost profile that survives contact with a finance team. That is the month this is.
One month from now you will have shipped a working AI application — retrieval, tools, agents, deployed — and be able to reason about cost, latency and failure like someone who has run one in production.
Where this points
Roles this program is aimed at
Generative AI / LLM Developer
₹7–38 LPA
NLP Engineer
₹7–38 LPA
MLOps Engineer
₹8–35 LPA
AI Application Engineer
₹6–35 LPA
Salary ranges are Indian market data for 2026 compiled from published third-party research, shown as context for the roles this program targets. They are not our students’ outcomes, not offers, and not a projection. We do not offer placement services and we make no job or salary guarantee.
Honest fit
Is this the right program for you?
This is for you if
- You write code today and are comfortable debugging your own work
- You want to move into an AI engineering role rather than improve your current one
- You can give weekday evenings plus Saturday labs for a month
- You want to finish with something deployed, not a notebook
This is not for you if
- You do not code. AI at Work is built for you instead, and costs less
- You want AI theory or research — this is an engineering program
- You cannot give the Saturday labs. That is where the project actually gets finished
Fees are non-refundable, so we would rather talk you out of the wrong program now than take your money and have you leave in week two. Read the refund policy.
Before you decide
The things people actually ask
Is a month really enough to move into AI engineering?
For someone who already codes, yes — because you are not learning to program, you are learning a specific stack. Sixty sessions of contact time plus four full lab days is enough to build and ship a real retrieval-and-agent system. It is not enough to make you a research scientist, and we do not claim it is.
₹75,000 is a lot. What makes it worth more than the student program?
Three things: it assumes you can code, so no time is spent on basics you already have; the Saturday labs are supervised build time, which is the expensive part of any program; and the material is current-stack engineering rather than foundations. If you do not already code, do not pay this — take AI at Work instead.
Can I do this alongside a full-time job?
It is demanding. Two hours every weekday evening plus a Saturday lab is roughly sixteen hours a week for a month. People do complete it while working, and sessions are recorded — but go in knowing the commitment rather than discovering it in week two.
Will I get a job at the end?
We do not offer placement services and we make no job or salary guarantee. What you leave with is a deployed project, the repository behind it, and the ability to explain the engineering decisions in it — which is what an AI engineering interview actually examines.
Still the right fit?
Tell us where you are starting from and we will say honestly whether this program suits you — including if the answer is that it does not.
Curriculum
20 evenings and 4 Saturday labs
- 01How these models actually behave — and prompting that survives production, not demos
- 02AI-assisted development — Claude Code and Cursor in a real codebase: context files, working from a spec, reviewing and debugging generated code
- 03Calling models from your own code — APIs, structured output, streaming, cost and latency
- 04Retrieval — giving a model your own documents, end to end, and why naive RAG returns the wrong passage
- 05Tools and agents — function calling, and an agent that completes a multi-step task
- 06Making it trustworthy — evaluating an AI feature instead of judging it by vibes, guardrails, handling hallucination
- 07Ship it — deploy, monitor and control cost, with the capstone built in the Saturday labs
This is the module outline, not a session-by-session syllabus. Order, depth and the specific tools used are adapted to each batch and to the trainer leading it.
FAQ
Questions about Applied Generative AI
Who is Applied Generative AI for?
Working IT professionals switching into AI careers. You should already be comfortable writing code.
How long is Applied Generative AI and how many hours is it?
One month, across 24 sessions in total — 20 weekday evenings and 4 Saturday labs, not 24 plus 4. Monday to Friday evenings, scheduled around working hours. Four full-day Saturday labs, ending in a Demo Day — in the Himayat Nagar classroom for the Hyderabad batch, on Google Meet for the online batch.
How much does Applied Generative AI cost?
₹75,000 in total. The fee is the same whether you attend the classroom batch in Hyderabad or the live online batch. They run separately, with the same curriculum, the same number of sessions and the same trainer time.
Can I attend this online from outside Hyderabad?
Yes. Online batches are live and instructor-led, not pre-recorded video, and students join from across India. Every session is recorded so you can catch up if you miss one.
What do I get at the end?
A certificate of completion, and more importantly the project work you produce during the program. Leave with a deployed project and the repository behind it.
Not sure this is the right one?
Tell us where you are starting from and what you want on the other side. We will tell you which program fits — including if the honest answer is none of them.