Learn Computer Stuff
Home / AI / Machine Learning / Conversational AI Developer
AI / Machine Learning

Conversational AI Developer

You build the assistant, and you make it behave when a user says something you never planned for.

No degree needed Approachable to start, hard to make reliable.


Can I actually do this?

Among the more open AI roles, because the hard modelling is done by someone else — the work is application engineering plus judgement about behaviour. Open without a degree. The realistic caution is that a demo is easy and a dependable assistant is not, so what distinguishes candidates is handling the failure cases rather than building the happy path.

Who it suits. People who enjoy the gap between what a user typed and what they meant.

Runway. Months if you can already build a web application.

Coming from another job?

Coming from web development? This is a shorter jump than the AI framing suggests — the model is an API and the rest is engineering you already do. Backend Developer Full-Stack Developer AI Software Engineer

Also advertised as

  • Chatbot Developer
  • LLM Application Developer
  • Assistant Engineer
  • Dialogue Engineer

The route

Four stations, in order. Each one is a thing you finish before the next matters.

  1. Station one

    Learn it free

    Only the best few, deliberately. Every one of these is free to use — the pill on each card says exactly what is and isn't free.

    Prompt engineering overview (Anthropic)

    Free to learn · no certificate

    How to actually steer a model, from a provider's own documentation. Free to read; using the API is metered.

    OWASP Top 10 for LLM Applications

    Free to learn · no certificate

    Prompt injection is not a curiosity — it is the standing vulnerability of everything you will build here. Free.

    MDN — Learn web development

    Free to learn · no certificate

    Most of this job is an ordinary application with a model in the middle. Free.

    See the full catalog in the explorer →

  2. Station two

    Attest strategically

    Nothing worth buying. Vendor courses in this area are frequently free and change faster than any certificate can track. A deployed assistant that handles bad input gracefully is the argument; a certificate about a model version that no longer exists is not.

    Nothing here is worth paying for

    No credential needed

    Nothing worth buying. Vendor courses in this area are frequently free and change faster than any certificate can track. A deployed assistant that handles bad input gracefully is the argument; a certificate about a model version that no longer exists is not.

  3. Station three

    Prove it

    A certificate says you passed a test. These say you can do the job.

    An assistant that refuses well

    Show it handling an out-of-scope, hostile or ambiguous request without inventing an answer. This is the difference between a demo and a product.

    An evaluation set you wrote

    A collection of inputs with expected behaviour, run automatically. Without one, you are changing prompts on vibes.

    A prompt injection you defended against

    Show the attack and your mitigation. Very few applicants can.

  4. Station four

    Get hired

    Search these exact titles

    • conversational AI developer
    • LLM application developer
    • chatbot developer
    • AI engineer

    Who hires for this. Customer support platforms, SaaS products adding assistants, agencies, and internal tools teams almost everywhere.

    A deployed assistant can be tried adversarially by an interviewer in about a minute, which is why robustness rather than polish is what carries. That is our reasoning about what is inspectable, not a verified hiring statistic.

    On salary

    We don't publish salary estimates. Numbers copied between blogs drift from reality, and a wrong number costs you real negotiating power. When we have a verified public source, it goes here with its date.


Where this route continues

· How we verify