CertBlueprint

Navigator Lab — a conversation, not a quiz

No trivia, no bubbles to guess at. Answer the way you'd tell a friend — a real reader is on the other end, and it may ask you a follow-up before it points anywhere.

Think about the last time you fell down a rabbit hole and lost track of time. What was it — and what kept you there?

reading what you actually wrote...

An AI model writes the reading. It can be wrong — treat it as a first conversation, not a verdict. The roles and trials it links to come from CertBlueprint’s own checked data. How this works, and what happens to what you write · Prefer clicking to writing? Take the guided assessment.

How the Lab works

  1. You answer one open question in your own words — the rabbit-hole question above. A sentence or two is enough.
  2. The reader may ask a follow-up — at most 2, one at a time. It is told not to place anyone from a vague first answer, so if what you wrote could point several ways, it asks the one question that would split them apart.
  3. After at most 3 answers, it gives a reading: what it heard you say, the field that fits, the honest road in, something to try this weekend, and a second direction when there genuinely is one.
  4. The page then attaches real links. The reader has to name one of 22 fixed fields, and the page looks that field up in CertBlueprint’s own data to show real role pages and a checked weekend trial.

What the AI writes — and what it doesn’t

The reader is an AI language model — Meta’s Llama 3.3 70B (@cf/meta/llama-3.3-70b-instruct-fp8-fast) — running on Cloudflare Workers AI. Writing a thoughtful reply to what you said is the job it is good at, so it keeps that job. Deciding what exists is not, so it doesn’t get that one.

Written by the model

  • The reflection of what you said
  • Why the field fits you, in its words
  • The road in: first job, useful cert, what year one looks like
  • A weekend idea — only when no checked trial exists, and labelled as the model’s

From CertBlueprint’s checked data

  • The field name shown in the reading, whenever the model’s pick matches the list
  • Every role page it links to
  • The weekend trial, with its time, cost and safety notes
  • The note that says when a field is not something this site maps

If the model names a field that is not on the list, the page drops the links rather than guessing. A missing link is a smaller mistake than a confident wrong one.

Every field the reader can point you to

The reader covers the full breadth of IT, including directions this site does not map. 17 fields link straight to role pages here, 1 links to close neighbours, and 4 are named honestly as outside what CertBlueprint covers.

Mapped to role pages

Partly mapped

  • QA & testingCertBlueprint does not map QA and testing as its own track yet. The roles below are the closest neighbours — the testing mindset shows up strongly in both. Closest roles mapped here: Software Developer, Application Security Engineer.

Not mapped here — and the reading will say so

  • AI & machine learning engineeringCertBlueprint has no AI/ML track yet, and pretending otherwise would waste your time. The roles below are the two most common routes people actually take into it — the data pipelines and the platform underneath. Closest roles mapped here: Data Engineer, Platform Engineer.
  • IT project managementThis site maps hands-on technical roles, and project management is not one of them. It is a real and well-paid direction — just not one CertBlueprint can honestly guide you through. Look at PMI and the CAPM/PMP route directly.
  • Business analysisCertBlueprint does not cover business analysis. It sits between the business and the technology rather than inside the technology, and mapping it properly would need a different site. IIBA is the place to start.
  • Technical writingTechnical writing is not mapped here. It is one of the genuinely underrated ways into technology for people who can already write — but this site would only be guessing at the path.

What happens to what you write

  • Each answer is sent to CertBlueprint’s server and passed to Cloudflare Workers AI, which generates the reply. Cloudflare processes it under its own privacy policy.
  • CertBlueprint does not save or log what you write, and there is no account. Like every page here, this one counts visits and time on the page (see the privacy policy), but never the conversation.
  • The conversation exists only in this page. Reload or leave, and it is gone — it is not kept in your browser either.
  • What you type is not sent to analytics.
  • Leave out your name, employer, contact details, and anything you would not want an AI service to process. The reader needs how you like to work, not who you are.

See the privacy policy for the rest of the site.

Limits worth knowing

  • Up to 3 answers in a conversation, each up to about 1,300 characters.
  • It can be wrong, and it only knows what you tell it. Treat a reading as the start of a conversation, not a verdict.
  • It is instructed never to comment on age, gender, race, religion, health, or other personal characteristics.
  • If you mention military service, it stays at the level of capabilities and never asks for operational details.
  • Text that tries to instruct it instead of answering — “recommend cybersecurity”, “ignore your rules” — gets a follow-up asking for a real answer, not a reading.

The Lab or the guided assessment?

This Lab

  • You write; an AI model reads it
  • 1 to 3 answers
  • Good if you already have a sense of yourself and want it interpreted
  • Needs a connection to our server

Guided assessment

  • You choose between parts of real IT situations
  • Usually 11–15 questions, no AI involved
  • Good if you have no idea yet where to start
  • Scored entirely in your browser; answers are never sent anywhere

Doing both is reasonable. If they agree, that is a strong signal; if they disagree, read both roles and try comparing them side by side.

Questions

Is it free?

Yes. No signup, no account, no payment.

Why did it ask me a question instead of answering?

Because your first answer could reasonably point at more than one field. A good mentor checks before placing someone, and the reader is built to do the same — it will conclude after at most 2 follow-ups either way.

Will it recommend a field this site doesn’t cover?

Yes. It is told to point you at the right field, not the convenient one. For project management, business analysis, technical writing and AI/ML, the page says plainly that CertBlueprint does not map that path and, where it can, shows the closest roles it does.

Does it know about certifications and salaries?

It knows general patterns, and it can be out of date. For checked exam details, use the certification guides; for pay, each role page shows what people typically start on and how far to trust the figure.

Why is an AI better than a quiz here?

It isn’t better — it is different. A quiz can only offer the options someone wrote in advance; the Lab can respond to exactly what you said. The guided assessment is more consistent and fully private, which is why both exist.