Back to index
Other28 days (live cohort)·$997 standard, $4,997 VIP with one-to-one coaching

AI Agent Bootcamp

3.3

A well-run cohort with genuine community energy and a credible instructor, wrapped in pricing and marketing that belong to a different category of product. The teaching is real. The $997 is hard to justify against what you can get free.

What We Liked

  • Tina Huang is a real practitioner with a real background, not a repackaged affiliate marketer
  • Cohort format with live sessions and accountability genuinely helps people who stall on self-paced material
  • Focus on multi-agent systems rather than generic prompt engineering is more current than most paid AI courses
  • Community and peer feedback are the strongest part of the product

What Could Be Better

  • $997 for 28 days is steep against free material from LangChain, Hugging Face and DeepLearning.AI that covers the same ground
  • The $4,997 VIP tier is the kind of pricing ladder that should make you cautious
  • Cohort dates are frequently listed as TBA, which makes planning awkward
  • Marketing leans on income and career outcomes more heavily than the curriculum warrants

Detailed review

Lonely Octopus is Tina Huang's company, and the AI Agent Bootcamp is its flagship. Huang is a former Meta data scientist with a pharmacology degree from Toronto and a computer science master's from Penn, and she built an audience north of a million subscribers on YouTube explaining data work clearly. That background matters, because the paid AI course space is thick with people whose only demonstrable skill is selling paid AI courses. She is not one of them.

The programme runs 28 days as a live cohort covering how to understand, build and deploy AI agents, with the emphasis on multi-agent systems where several specialised agents coordinate on a task. That is the right thing to be teaching in 2026. Most paid AI courses are still relitigating prompt engineering, which stopped being a scarce skill some time ago. Teaching orchestration, tool use and deployment is closer to what people actually need.

The cohort model is the real product here, and I want to be fair about it. There is a large population of people who buy a Udemy course, watch two hours, and never open it again. For them a fixed 28 day window with live sessions, peers and someone noticing whether they showed up is worth money. The community is consistently the thing participants rate highest, which tracks with what the format is genuinely good at.

Then there is the price. $997 for four weeks, with a $4,997 tier attached that adds one-to-one coaching. I have a general rule that a pricing ladder with a five times multiplier at the top is a marketing structure before it is an educational one, and that rule has not let me down often. The comparison that matters is not other bootcamps.

It is the free tier of this exact subject. Hugging Face runs a free agents course. LangChain and LlamaIndex both publish extensive free tutorials. DeepLearning.AI has multiple free short courses on agent frameworks.

Anthropic and OpenAI both publish free cookbooks on tool use and orchestration. Someone who works through that material for 28 days with the same discipline a cohort would impose on them will end up in a similar place and be $997 better off. The honest framing is that you are paying for structure, deadlines and peers, not for information. That is a real product and some people should buy it.

Just be clear with yourself about which of those two things you are short of. The other wrinkle is scheduling. Cohort dates are often listed as TBA with a notification form, which is a reasonable way to run a business and an annoying way to plan a month of your life around something. Three point three.

Credible instructor, sensible curriculum, genuinely useful format for people who need external structure, priced well above what the underlying material justifies.

[ final ]

The verdict.

Worth it if you have tried to learn agents on your own, bounced off, and know that a deadline and a cohort is what actually makes you finish. Not worth it if you are disciplined, because the free alternatives are genuinely good now.