EXECUTIVE POST GRADUATE CERTIFICATE PROGRAMME
Generative AI and Agentic AI
Stop prompting AI. Start building systems that act.
4 Months
TBC
90 Live Hours
8 to 9 months, 180 hours
FORMAT
Weekends, live online
Start Date: 15th October, 2026
The next AI skill is notprompting. It is building systems that act
AI has moved beyond answering questions. The next generation of AI engineers will build systems that can plan, use tools, call APIs, retrieve information, correct their own mistakes, and complete complex tasks with minimal human intervention.
That means the real engineering challenge is no longer just the model. It is the system around it, the agents, workflows, retrieval, evaluation, and infrastructure that make AI reliable in the real world.
“We believe that, in 2025, we may see the firstAI agents ‘join the workforce’ and materiallychange the output of companies.”
Sam Altman
CEO
The market is already hiring for Generative & Agentic AI
Build the skills to move from using AI models to engineering systems that plan, act, and ship.
Generative AI Engineer
Agentic AI Specialist
ML Engineer (LLM)
AI Solutions Architect
INDICATIVE COMP · INDIA
₹15 LPA–₹50 LPA
Across generative and Agentic AI engineering roles at 3-5 years of experience.
INDICATIVE COMP · INDIA
$127K–$375K
Across the US, UK and Gulf markets.
*Salary ranges are indicative market ranges compiled from Glassdoor, LinkedIn and NASSCOM job-posting data; not a programme placement or salary guarantee.
Who is this for?
Built for engineers and technical professionals who want to move beyond using AI tools and start building the systems behind them.
Technical background
Engineering or technology graduates from disciplines such as CS, IT and ECE.
Python & Maths
Foundational Python along with mathematics, statistics and basic AI/ML.
50%+ UG
A minimum 50% undergraduate score.
PROGRAMME HIGHLIGHTS
Built to be practical
From AI Fundamentals to Agentic Systems
180 hrs
across 8 to 9 months of weekends
16
modules, five stages
27+ hrs
end-to-end agentic capstone
CURRICULUM
Sixteen modules, one working system
180 hours across five stages — foundations,generative AI, retrieval and fine-tuning, agentic systems, and production deployment.
M0 to M1 · Foundations
M0 to M1 · Foundations
Python recall, AI, ML, and deep learning, neural networks, CNNs, and backpropagation.You walk out able to: work fluently across the classical stack, so nothing in the LLM modules rests on a gap.
M2 to M3 · NLP and Embeddings
M2 to M3 · NLP and Embeddings
Transformers, BERT, vector databases, semantic search, FAISS.You walk out able to: build with transformer architectures and run semantic retrieval over a vector store you set up yourself.
M4 to M5 · Generative AI
M4 to M5 · Generative AI
GANs, VAEs, diffusion models, LLM architecture, open-source LLMs.
You walk out able to: build across text, image, and multimodal generation, and work inside an open-source LLM rather than around it.
M6 to M7 · Prompt Engineering and LangChain
M6 to M7 · Prompt Engineering and LangChain
Prompt strategies, structured JSON output, LangChain chains, memory, introduction to RAG.
You walk out able to: enforce reliable structured output and chain memory-aware pipelines.
M8 to M9 · RAG Systems
M8 to M9 · RAG Systems
Basic RAG, Advanced CRAG, Graph RAG on Neo4j, RAGAS evaluation.
You walk out able to: ship production retrieval and prove it works with a real evaluation harness.
M10 · Fine-Tuning LLMs
M10 · Fine-Tuning LLMs
LoRA, QLoRA, instruction tuning, PEFT, LM Eval Harness.
You walk out able to: fine-tune efficiently on constrained hardware and benchmark the result.
M11 · Agentic AI and Multi-Agent Systems
M11 · Agentic AI and Multi-Agent Systems
LangGraph, Microsoft AutoGen, CrewAI, Agentic RAG, multi-agent systems.
You walk out able to: build agent teams that plan, delegate, and recover from their own failures.
M12 to M14 · LLMOps and Deployment
M12 to M14 · LLMOps and Deployment
MLFlow, Docker, Kubernetes, FastAPI, LangSmith, low-code and vibe coding.
You walk out able to: deploy, monitor, trace, and scale an AI system in production.
M15 · Capstone
M15 · Capstone
End-to-end agentic build, deployment, and evaluation by IIT Roorkee faculty.
You walk out able to: run one problem from architecture to a live, monitored system and defend every choice in it.
By the end, you'll be able to
The role of the AI engineer is changing
AI has moved beyond making predictions or generating responses. The new challenge is designing systems that can plan, act, retrieve, verify, recover and run in production.
YESTERDAY
The focus was the model.
One model, judged on its predictions.
TODAY
It's the system around it.
Plan
ACT
RETRIEVE
About IIT Roorke
Founded in 1847, IIT Roorkee is one of India’s oldest and most prestigious institutions, with a legacy of academic leadership and technological innovation. As a pioneer in interdisciplinary education, IIT Roorkee has been at the forefront of engineering, data science, AI, and management education.
Its executive education programs are designed to equip working professionals with industry-aligned, future-forward skill sets, combining academic rigor with hands-on learning and strategicthinking.
WHERE THIS TAKES YOU
Roles this supports
Generative AI Engineer
₹10–20 LPA
Agentic AI Specialist
₹12–25 LPA
ML Engineer (LLM)
₹10–22 LPA
AI Solutions Architect
₹20–45 LPA
CAREER ASSISTANCE
1:1
Mentor and career sessions
Portfolio
Capstone build as proof of work
Referrals
Hiring partner introductions
Mocks
AI PM interview
Our students are acing it
They are working at companies that are a dream for most
Fee Structure
Programme fee
₹1.4 Lakhs + GST
EMI options available through our NBFC partners. Speak to an advisor for the current plans.
PAYMENT
Application fee
₹5,000
Non-refundable, adjusted in final fee.
Programme fee
₹1,40,000 + GST
One-shot, or EMI via NBFC partners.
Advanced Certificate from IIT Roorkee in Generative & Agentic AI Engineering.
A faculty-led curriculum spanning LLMs, RAG, Agentic AI, multi-agent systems, and AI deployment.
A 27+ hour capstone to architect, deploy, evaluate, and defend an end-to-end Agentic AI system.
Led by the Futurense Leadership Council (FLC)
A collective of CXOs, AI leaders, and digital transformation heads from global and Fortune 500 companies shaping the AI-native workforce.
Frequently Asked Questions
We know you might have some questions before getting started in our platform
What is the Forward Deployed AI Product Management programme?
An advanced certificate programme from IITM Pravartak for product professionals who want to work AI-native. It covers building and evaluating AI and agentic products, managing the AI product lifecycle, and owning end-to-end delivery. Four months, weekends, 90 hours, one capstone.
What does "forward deployed" mean in a product role?
It describes how you work, not where you sit. You own a product outcome end-to-end: align stakeholders, gather requirements, prototype quickly, work alongside engineering and design, and drive returns quickly. No travel or client-site posting is involved. You own the outcome, not just the backlog.
Is the programme lecture-based or hands-on?
Hands-on throughout. Every phase produces a working output: prototypes, PRDs, agent-ready specs, and evaluation suites. Concepts are introduced in service of what you are building rather than as standalone theory.
Who teaches the programme?
IIT & IIM faculty alongside experienced industry practitioners. Practitioners carry much of the hands-on teaching and both expert masterclasses. Concepts are anchored academically, while the Tech part is taught by IIT Faculty & the Consulting is covered by the IIM faculty.
What are the eligibility requirements?
A graduate degree in any discipline. Field of study matters far less than whether you already think in terms of users, problems, and outcomes.
Is coding knowledge required?
No. You direct AI through prompting, not production code. Prototypes are built with AI tools, which is a different craft from software engineering and requires no programming background. Any graduate discipline works.
What is the admission process?
Three steps. Submit an application with your details and your interest in the programme. Clear a short, non-technical qualifying test that assesses your skills. Then confirm your seat with payment to secure a place in the upcoming cohort.
Is there a qualifying test, and is it technical?
Yes, there is a short qualifying test as part of admissions. It is non-technical and does not assess coding, so no programming background is needed. It is designed to check whether you are a fit for the programme's pace and level.
What is the programme fee?
The programme fee is ₹49,999, plus 18% GST as applicable as applicable.
Is the campus immersion included in the fee?
No. The two-day campus immersion at IIT Madras Research Park is optional and charged separately. Travel, accommodation, and all related expenses are borne by the participant. The immersion is not required to complete the programme or receive the certificate.
Are there any other costs?
No. The programme uses only free, free-to-try, or open-source tools, with no paid subscriptions and no proprietary sandbox to purchase. This is deliberate: real product work is done with exactly these tools.
Are EMI or financing options available?
Yes, EMI options are available through partner NBFCs.
What is the refund policy?
If a student withdraws before program commencement, fees are refundable, excluding registration fee, GST, and subvention charges. No refund is applicable after the program begins.
What is the total duration and time commitment?
About four months, 90 hours in total: 72 hours of taught modules, 8 expert masterclasses, and 10 hours of the capstone. The 10 capstone hours are your own project time, not additional classes.
When are classes held?
Teaching runs on weekends, so the programme fits around a full-time job.
Is the programme fully online?
Delivery is online. An optional two-day campus immersion at IIT Madras Research Park is available for participants who want the in-person experience. It is not required to complete the programme or receive the certificate.
How is the programme structured across the four months?
Seven modules run in sequence across weekends, grouped into five phases: the forward-deployed mindset, understanding LLMs and agents, building and prototyping, proving and engineering, and forward-deployed delivery. Masterclasses sit inside that flow as applied checkpoints. The capstone comes near the end.
What is the capstone project?
An end-to-end agentic product build. You take an idea through the full lifecycle: discovery, prototype, evaluations, a working loop, specs, and a ship plan. It is the only individual submission and the only assessment in the programme.
Is the capstone mentored?
Yes. It runs as a mentored end-to-end build, so you have guidance through the full lifecycle rather than submitting a project unsupported.
Are there exams?
No exams. The capstone is the single assessment. Completing it completes the programme.
What are the expert masterclasses?
Expert masterclasses consist of live sessions where a practising expert builds a real AI product case in front of the cohort, so you see how the work is actually done. Cases are drawn from current industry problems and refreshed as the market moves.
What certificate is awarded, and who awards it?
An advanced certificate from IITM Pravartak, a Technology Innovation Hub of IIT Madras, is awarded upon completing the programme. Futurense is the admissions partner.
Is this a certificate or a degree?
An advanced certificate programme, not a degree. It is designed as a focused professional upskill for working product professionals rather than an academic qualification.
How does this programme support career growth?
Product roles are shifting toward people who can build and ship with AI, not only specify features. You finish able to do that work, with an IITM Pravartak credential and a portfolio of real builds to show for it.
Will the content stay relevant as AI changes?
The programme is built on durable skills: how to evaluate an AI product, when agents or loops are justified, and how to run an agentic lifecycle. Cases and masterclasses are refreshed to stay current, so the way of working outlasts individual tools.