Architect — AI Fluency
Twelve weeks that take you from writing prompts to running autonomous systems in production.
₹1,49,999
EnrollArchitect is the whole arc. Twelve weeks, twenty-four live sessions, three finished systems — and a genuinely different relationship with the technology by the end of it.
You begin where everyone begins: prompts, models, and a working agent by the close of the first stage. The second stage makes that agent capable — retrieval, tools, MCP, memory. The third is the part most people never reach: multi-step reasoning, several agents working together, evaluations that tell you whether any of it is actually good, guardrails that keep an autonomous system inside its brief, and finally deployment, so the thing keeps running after you close your laptop.
Each stage ends with a project built across two guided sessions. You finish with three working systems, not three abandoned experiments.
Exam access is included rather than bought separately. Complete Architect and you are offered the Upfyn AI Internship on the live Upfyn project, and internship exam retakes are included for as long as your community access lasts.
Sessions are taught live on fixed dates, two a week, in a small batch, and your group starts on the date shown at checkout. Covered by a 15-day refund — no questions asked.
Why start now
AI that answers questions is everywhere. AI that runs on its own, checked, safe and deployed, is rare, and the people who can build it are rarer still. India alone could need more than 2.3 million people in AI roles by 2027, with about 1.2 million to fill them. Architect trains you for the scarce side of that gap.
What you will learn
- Build practical AI agents from a standing start
- Create grounded workflows with RAG, tools, MCP and memory
- Design multi-step and multi-agent systems that stay observable
- Evaluate what you build and keep it inside safe limits
- Deploy an autonomous workflow that runs without you
- Finish three portfolio-ready AI systems
Tools you will work with
You practise on the apps your work already runs on. Upfyn is built to connect with all of them, so what you build in class works where you work.
AI assistants
- ChatGPT
- Claude
- Gemini
- GitHub Copilot
- Mistral
- DeepSeek
- Meta AI
- Ollama
- Hugging Face
- Cursor
- Replit
Docs & sheets
- Notion
Chat & support
- Slack
Projects & people
- Linear
Automation
- n8n
- Make
- Zapier
- MCP
Build & analytics
- GitHub
- GitLab
- Python
- JavaScript
- TypeScript
- Jupyter
- Docker
- Kubernetes
- Vercel
- Supabase
- PostgreSQL
- MongoDB
- Firebase
- Google Cloud
- awsAWS
- LangChain
- Snowflake
- Databricks
- BigQuery
- Looker
- BIPower BI
- TTableau
Who it is for
- You want to build AI systems that run in production, not just prototypes
- You lead a team or a product and need to direct AI work with confidence
- You are aiming for the internship on the live Upfyn project
- You are ready to commit twelve weeks to going all the way
Why this course is the right fit
The whole arc in one path. Prompts to agents, agents to connected systems, then multi-agent designs, evaluations, guardrails and deployment.
Three systems, not three experiments. Each month ends with a project built across two guided sessions. You finish with three working systems in your portfolio.
Evidence that it works. Evaluations and guardrails are taught, not skipped, so you can show a system is good instead of saying so.
The route into real work. Exam access is included, and when you complete the cohort you are offered the internship on the live Upfyn project.
What's included
24 live sessions over 12 weeks
Two a week, taught by a mentor in a small batch. Every session is recorded if you miss one.
One-to-one doubt sessions
Ask for time with a mentor whenever you are stuck, for as long as your cohort runs.
Upfyn AI credits
Credits come with your course, so you practise on real AI from the first session.
A year in the community
One year of community access with this cohort. Join three cohorts and it is yours for life.
Certificate, with your portfolio on it
Complete the course and you get a certificate that shows the portfolio you built: the projects you shipped, not the hours you sat through.
AI Fluency exam included
Exam access comes with the course, not as a separate purchase.
Internship included
Complete your cohort and you are offered the internship on the live Upfyn project.
Exam retakes included
Not through to the internship by the end of your cohort? Keep re-attempting the internship exam after it ends, for as long as your community access lasts.
15-day refund
No questions asked, 15 days or before.
Syllabus
Month 1 · 8 live sessions
Foundations, and your first agent
- 1AI Introduction
- 2Basics of LLM
- 3AI Prompt Engineering
- 4AI Hygiene
- 5Automation, AI agents and agentic work
- 6Action Planning
- 7First Line Agent — project, day 1
- 8First Line Agent — project, day 2
Project: First Line Agent. You ship a working agent, from a standing start.
Month 2 · 8 live sessions
How agents actually work
- 9Anatomy of Agents
- 10LLM Preprocessing
- 11Vector Databases
- 12RAG Fundamentals
- 13API Wrappers, Tools, MCP
- 14Agent Memory
- 15Agentic Workflow — project, day 1
- 16Agentic Workflow — project, day 2
Project: Agentic Workflow. You build an agent that remembers, retrieves and uses real tools.
Month 3 · 8 live sessions
Systems that run without you
- 17Multi-step Reasoning
- 18Multi-Agent Systems
- 19Agentic Frameworks
- 20Agentic Evaluations
- 21Safety and Guardrails
- 22Custom Toolkits
- 23Autonomous Loop — project, day 1
- 24Optimise and Deploy — project, day 2
Project: Autonomous Loop, deployed. You deploy something autonomous, evaluated and running.
What you walk away with
- Three portfolio-ready AI systems, including an autonomous loop
- The judgement to design, evaluate and deploy multi-agent systems
- The AI Fluency exam, included
- The internship on the live Upfyn project, offered when you complete
Questions before you join
What people ask about Architect before they start, answered plainly.
Twelve weeks is a big commitment.
It is two live sessions a week, recorded if you miss one, with one-to-one doubt sessions whenever you are stuck, and it is designed to fit around a full-time job. Twelve weeks from now will arrive either way; Architect decides what you will be able to build when it does.
Is it worth the step up in price?
Architect is the only course that takes you all the way to deployed, evaluated systems, includes the exam, and includes the internship when you complete it. Workers with AI skills already earn an average 62% more, according to PwC's 2026 barometer, and Architect is where the scarcest of those skills are taught.
Do I need to be an engineer for multi-agent systems?
No. The first two months build every foundation Architect relies on. By the time you reach multi-agent design you have already built and connected agents yourself, with a mentor reviewing each project.
Is the internship really included?
Yes. Complete your Architect cohort and you are offered the internship on the live Upfyn project. Exam retakes are included too: if you are not through by the end of your cohort, you can keep attempting the internship exam after it ends, while your community access lasts.
Ready to start?
Twelve weeks and twenty-four live sessions — three systems that run, plus exam access and the internship included.
₹1,49,999
EnrollRelated reading
All postsAI Careers · 23 Sept 2026
AI skills pay more, and the premium keeps growing
Workers with AI skills earn an average 62% wage premium, per PwC's 2026 barometer. Here is what the data shows, and which kinds of AI skill are paid for.
AI Skills Gap · 25 Sept 2026
India's AI skills gap: 2.3 million AI jobs, 1.2 million people
India could have more than 2.3 million AI job openings by 2027 and about 1.2 million people to fill them. What the gap is, and why it will not close by itself.
AI Careers · 21 Sept 2026
Employers now prefer AI skills over degrees
40% of Indian employers now prefer demonstrable AI skills over a degree. What that means, and what proof of AI skill looks like to someone hiring.