GYANHQ SUPER 30 · AI & FORWARD DEPLOYED ENGINEER (FDE) TRAINING
Build AI that works beyond the demo.
A mentor-led cohort for students, engineers and founders ready to turn AI knowledge into a working system.
Personally reviewed. Selected candidates meet us one-to-one.
- 30
- builders per cohort
- 3 months
- 12-week learning journey
- 1 capstone
- build, evaluate & deploy
From a spark of curiosity
to something you can ship.
Agents · Retrieval · Identity · Evaluation · Deployment
01 / YOUR STARTING POINT
Your experience.
Your next step.
Different starting points. The same focus on building something that works.
ONE SHARED OUTCOME
A build you can explain.
A system you can improve.
Work through the decisions behind a useful AI application: what to retrieve, which tools to trust, how to evaluate it and how to deploy it.
Choose a starting point to explore what you could work toward.
Request Super 30 Invite02 / FROM IDEA TO SYSTEM
See what the work
looks like.
A sample Enterprise RAG Assistant, based on one of the programme’s capstone briefs.
An assistant that knows when to answer.
Help a team find a policy, cite its source and respect access boundaries.
THE BRIEF
The answer is somewhere.
Your team needs it now.
Policies are scattered across documents. Build an assistant that retrieves relevant information, provides a grounded answer and links back to the source.
- Retrieve relevant passages
- Cite the document behind the answer
- Respect the reader’s permissions
- Say when the evidence is missing
Give every answer a traceable path.
- 01QuestionIdentify the user and request.
- 02RetrieveFind permitted source passages.
- 03GroundAnswer using the evidence.
- 04EvaluateCheck quality and trace the result.
The build also needs ingestion, identity checks, useful logs and a deployment plan. Each decision becomes part of your engineering work.
Ask the sample handbook.
Illustrative questions and prepared answers from fictional documents.
Submit an equipment request to your team lead before ordering. The team lead reviews and approves the request.
Source: Fictional equipment handbook · Section 2
Inspect the sample sources
Equipment handbook, Section 2: “Submit an equipment request to your team lead before ordering. Your team lead reviews and approves the request.”
Source coverage: No travel allowance policy is included.
Access rule: Payroll documents require payroll-team access. The sample reader does not have that access.
A fluent answer is only the beginning.
Does each factual answer agree with the cited passage?
Does restricted information stay restricted?
Does the assistant avoid inventing missing policies?
Can you inspect latency, cost, failures and useful traces?
These are evaluation criteria for your build, not claimed performance results.
03 / LEARN WITH A PRACTITIONER
Meet Inder Chauhan.
Enterprise AI & platform engineering instructor
Explore the reasoning behind architecture, production deployment and reliability. The programme brings mentor reviews into the work, from your early decisions to your capstone.
Experience & teaching approach04 / THREE MONTHS OF BUILDING
A clear path.
Work you can show.
A 12-week learning journey. Two teaching classes and a 2.5-hour hands-on lab each week build toward your capstone.
- WEEKS 1–2
Build the foundations.
Engineering fundamentals, LLM applications and predictable outputs.
- WEEKS 3–6
Give AI context & tools.
Retrieval, citations, evaluation and controlled agent workflows.
- WEEKS 7–10
Make it work in production.
MCP integration, cloud deployment, observability and safety.
- WEEKS 11–12
Deliver the capstone.
Discover a business problem, deploy your solution and explain the tradeoffs.
A SMALL TASTE OF THE THINKING
The source is missing.
What should AI do?
A user asks your policy assistant for a travel allowance. None of the available documents gives an amount.
Choose an answer to explore the tradeoff. No signup needed.
05 / BEFORE YOU REQUEST AN INVITE
A few useful answers.
Make an informed decision about your next step.
A conversation,
before a commitment.
- Send your invite request.
- We review your background and goals.
- Selected candidates join a 30-minute one-to-one session.
We discuss the start date, fees, weekly commitment and live-session timezone in that conversation.
Who is Super 30 for?
Students, engineers and founders who want to build AI systems beyond a demo. Share your background and goals so we can discuss whether the cohort fits your starting point.
How long is the programme?
Three months: 12 weeks, with two teaching classes and one 2.5-hour hands-on lab each week. You can explore the complete curriculum on the Program page.
What are the fees, dates and weekly hours?
The weekly format is two teaching classes plus a 2.5-hour lab. Fees, dates, teaching-class timings, independent study expectations and live-session timezone are confirmed in a 30-minute one-to-one session with selected candidates.
Does requesting an invite confirm my place?
No. Every request is reviewed. Selected candidates are invited to a conversation before the next steps are agreed.
When will I hear back?
We aim to respond within 72 hours, in line with the application confirmation. Keep an eye on the email address you provide.
What will I build?
One deployed capstone. Programme examples include a document assistant, an agent workflow or a support assistant. Your work includes evaluation, integration and explaining the engineering decisions.
YOUR NEXT CHAPTER STARTS WITH A CONVERSATION
Bring your curiosity.
Build something useful.
Three months. A cohort of 30. One capstone to put your learning to work.
Request Super 30 InviteRequests are reviewed individually. An invite request does not guarantee a place.
Explore the GyanHQ ecosystem
Payments, enterprise workflows and AI assistance. Explore the products behind the wider GyanHQ ecosystem.