Join DECIL Cohort 2
- One in-person class a week for five weeks, livestreamed at the same time — four modules plus Demo Day — hands-on rather than lecture. You build during the program, not after it.
- Go from a raw idea to a working prototype, workflow or pitch you can actually show.
- Set up an agent crew that does your research, build, content and operations work — and a business model that pays for it.
- Present at a full Demo Day, hosted and judged by the instructor. Top projects are considered for mentorship, incubation and investor introductions.
- One seminar hosted by DECIL rounds out the program.
Your instructor
Nathan Wang, PhD
Nathan Wang is an AI founder, full-stack developer and US engineering PhD with 15 years of hands-on work across hardware, software and AI at ASM, Lam Research, Apple and Meta. At ASM and Lam Research he developed thin-film deposition and etch processes for advanced semiconductor nodes; at Apple he helped ship iPhone X through iPhone 14, AirPods Pro and Mac Pro; at Meta he was responsible for AR glasses, AI agent integration and the wearables operating system. Today he is founder and CEO of Ipsion AI, where he builds and runs the AgentOS / ClawBot multi-agent system on his own as a full-stack developer, and through Soul AI he delivers hands-on AI-native training and transformation consulting for individuals and companies. He teaches AI on Maven, Udemy, LinkedIn Learning and more: 24 courses, 50,000+ students in over 50 countries. His conviction: the value of AI isn't how much headcount and cost it saves. It's giving one person a whole team, amplifying a single builder into the force of a company of dozens, with the nerve to take on the giants.
How the five weeks run
Every week is hands-on with ClawBot / AgentOS, the multi-agent system built by Ipsion AI.
OPC Foundation
Product × System × Business — the three things a one-person company has to hold at once, and how AI changes each of them.
System — set up your agent crew
Stand up the agent team that does your research, content, operations and engineering work.
Product — build what creates real value
Turn the crew into an agent product someone actually wants, and get it in front of a first user.
Business — revenue is all you need
Pricing, distribution and the shortest path from a working prototype to the first dollar.
Demo Day — show the world your OPC
Present live, hosted and judged by the instructor, with feedback on every project. Strong projects go to mentors and investors.
Seminar
One additional seminar hosted by DECIL alongside the program. Speakers, format and date announced by DECIL.
Agent-engineering craft — harness engineering, loop engineering, memory and skill design — isn't a single week. It runs through all five, revisited from the product, system and business angle each time.
What you'll learn
AI-native mindset
How AI changes the way one person builds companies, products and services.
Agent crew & harness
Stand up a crew of agents, and engineer the harness and loop that make them reliable.
Memory & skill design
Give agents memory that compounds, and package what works as reusable skills.
Product judgment
Decide what is actually worth building, then get it in front of a first real user.
Revenue
Pricing, distribution and the shortest path from working prototype to first dollar.
Local AI in practice
AMD Ryzen AI demonstrated live by the instructor — where local compute beats the cloud, and where it doesn't.
One seminar, outside the five weekly modules
DECIL hosts one seminar alongside the program. The Cohort 2 speaker, format and date are set and announced by DECIL.
- Cohort 1's seminar was Physical AI with Prof. Allen Yang, UC Berkeley. EECS faculty and Executive Director of the FHL Vive Center for Enhanced Reality; 100+ technical papers and 20+ US patents across computer vision, autonomous systems and immersive interfaces. That is the level DECIL books to — the Cohort 2 speaker is not announced yet.
Who it's for
No prior startup experience required. Curiosity, consistency and a willingness to build matter more.
University students
Interested in AI, startups, product, engineering, business or investing.
Early builders
Exploring AI-native venture ideas and wanting to move fast with fewer resources.
Side-project builders with a day job
Keeping your job while you start a company or side business with an agent crew doing the heavy lifting.
Researchers
Technical students and academics interested in commercialization and venture creation.
Creators & operators
Consultants and creators who want AI leverage and intend to operate solo.
Before you register
How much time does it take?
Five weekly sessions — four modules of about 1.5 hours each and a 2–3 hour Demo Day — plus roughly 5–10 hours a week of your own building, and one seminar hosted by DECIL. This is a builder program, not a lecture series: the people who get the most out of it are the ones who ship something between sessions.
Do I need to bring my own laptop?
Yes. Every participant is expected to bring a personal computer to every session. AMD hardware is used for live demonstrations by the instructor and is not distributed to participants.
Do I need a technical background?
No. You do need to be comfortable experimenting with tools and willing to iterate. The program is hands-on, so you'll be building either way.
In person or livestream?
Each week is one in-person class, livestreamed at the same time for anyone who can't be in the room. Same content and recordings either way. In-person capacity is limited and allocated first-come, first-served.
What happens after I pay?
You'll get a confirmation email from DECIL. Registration, payment and participant support for this program are run by DECIL, not by AI Superpower.