AI certifications explained: What employers value

Keith Shaw: There's been a surge in the number of AI certifications — from cloud providers and training companies to universities and boot camps. There’s no shortage of programs aiming to prepare people for the age of AI. But how do you know which ones matter?

And could they already be outdated by the time you finish the course? On this episode of Today in Tech, we're digging into what's real, what's hype, and where AI certifications go next.

Keith: Hi everybody, welcome to Today in Tech. I'm Keith Shaw. Joining me on the show today is Sydnee Mayers, Product Lead for AI at Cribl. Welcome to the show, Sydnee. Sydnee Mayers: Thanks, Keith, for having me. Keith: So let’s jump right into it.

You're in the space where you're evaluating people with AI skills, so you’ve probably explored a lot of different AI certifications. When did you first notice the explosion of different types of certifications? Was it right at the start, or did it come later after generative AI took off?

Sydnee: There was definitely a lag. ChatGPT put generative AI on the map in a way people hadn’t seen before — but initially, it was mostly consumers experimenting with it. Businesses and professionals hadn’t yet realized the immense value it could bring to day-to-day work.

I started to notice more interest in certifications around mid to late 2023. At that point, ChatGPT had been out for a while, and people began saying, “Wow, this AI stuff is legit. I need to skill up.”

Keith: When you look at the current landscape, how varied are these certifications? Can you give us examples of what you're seeing? Sydnee: There are two big camps. First, professional certifications — from Amazon, Google, Microsoft, even NVIDIA. Then there are academic certifications from universities like Harvard, Stanford, and MIT.

Both have value, but I see more professionals going the vendor route. They’re typically faster to complete, sometimes free, and often more accessible — especially if offered by your employer.

The vendor-driven ones — like Microsoft, Amazon, and Google — dominate what I see on LinkedIn, on resumes, and in hiring conversations. Influencer-backed programs, like Andrew Ng’s, are also popular. Interestingly, we haven't seen the same level of certification from OpenAI or Anthropic.

They offer great learning content, but not much in the way of certification exams. I suspect it’s because of their partnerships with Microsoft and Amazon, who already have well-established certification pathways.

Keith: That makes sense. Is the main difference between taking a course and getting a certification the testing component? Sydnee: Partly. Courses can range from weekend overviews to more in-depth programs, but certifications often involve a structured pathway — from fundamentals to specialized tracks like data engineering.

Certifications offer more rigor, structure, and validation. When I see a certification on a resume, I know that person had to prove something. Keith: Right — it’s more than just a certificate of attendance. Sydnee: Exactly.

Some courses are more like “click-through” training — not really something you’d put on a resume and call yourself “certified.” That’s where certifications stand out.

Keith: What problems are these certifications aiming to solve? Sydnee: Most begin with foundational knowledge — terminology, key concepts, the difference between generative and traditional AI. They’re democratizing access and building fluency.

Of course, they also introduce learners to the providers’ platforms and services — which is a smart branding move. Role-based certifications go deeper, teaching hands-on skills, architecture, and how to build using those services — from vector databases to chaining LLMs. So it's both platform exposure and skill-building.

Keith: You brought up an interesting point earlier — that traditional developer skills might not translate well into working with AI. Can you explain? Sydnee: Sure. Generative AI is non-deterministic — the same input can yield different outputs. Traditional development is deterministic.

Developers are used to two plus two always equaling four. So when outputs vary, it’s disorienting. They’re not used to designing systems where answers change slightly every time.

Front-end developers especially struggle with how to present non-deterministic results to users — how to convey that a different answer doesn’t necessarily mean an incorrect one.

Keith: Is part of that challenge due to unfamiliarity with prompt engineering? Sydnee: Definitely. Prompt engineering isn’t something most developers were trained in. Small wording tweaks — even switching “a” to “the” — can lead to different results. Developers need to get comfortable with that kind of linguistic nuance.

I often say LLMs are like English majors — they respond best to well-crafted language. Developers are used to structured, mathematical thinking. It’s a big shift.

Keith: So how do you help developers who are struggling? Sydnee: I usually suggest learning from communities — GitHub, Hugging Face — to see real-world prompting examples. Many providers also publish excellent prompt guides. And yes, certification courses often cover this too.

Keith: From a hiring perspective, do certifications matter more for recruiting or upskilling? Sydnee: Both. When I see a certification on a resume, it changes how I interview. I can assume we share a baseline understanding, which lets us dive deeper into architecture or problem-solving. It’s much more productive.

For internal upskilling, certifications help standardize learning across different departments and roles. They offer fundamentals, then specialized tracks — sales, marketing, engineering — which companies can deploy more easily than creating in-house training from scratch.

Keith: But do these certifications truly translate into real-world skills? Sydnee: That’s a challenge across all certifications. Just because someone passed a test doesn’t mean they can apply the knowledge. That’s why we’re seeing more practical hiring processes — live coding, case studies, architectural challenges.

Certifications help you assess a candidate’s vocabulary and framework, but you still need real-world application to know if they can do the job — and how quickly.

Keith: From the job seeker’s side, do people pursue certifications just to check a box or actually learn? Sydnee: I’d say 70% genuinely want to learn and apply the knowledge. The other 30% just want “AI” on their resume. That’s fair, given market demand.

The bigger issue is alignment — developers often get a fundamentals cert, which may not equip them for hands-on work. And for non-technical roles like sales or marketing, there just aren’t enough certifications yet tailored to their needs.

Keith: Is lack of standardization a problem? Sydnee: Absolutely. Even within “fundamentals,” each provider has a different approach. And AI is a broad field — research, development, consumer usage — so it’s hard to standardize. But even basic categorization (e.g., for builders, consumers, or platform users) would help a lot.

Keith: What about the pace of change? Could certifications become obsolete in a year? Sydnee: That’s already happening. In the deep technical space, things evolve fast — new frameworks, deprecated tools. It’s hard for certifications to keep up.

But the fundamentals are more stable — LLMs, prompt engineering, terminology — those will stay relevant longer.

Keith: Do you think AI certifications will become foundational like cloud certs, or stay fragmented? Sydnee: I don’t see full standardization happening. The ecosystem is too fragmented. I think fundamentals will become baseline knowledge — everyone will be expected to know them. But beyond that, specialization will vary a lot.

Keith: Are fairness, bias, or compliance issues being covered? Sydnee: Most certifications mention responsible AI in the fundamentals. But deeper content around bias, prompt injection, and AI security is still lacking. Hopefully that evolves — these are critical topics.

At my company, we always start AI training with responsible and ethical use before diving into tech.

Keith: Can certifications close the AI skills gap — or will they widen it? Sydnee: Certifications help, but staying current is about more than courses. You have to use AI regularly, stay engaged with news, and get hands-on.

A lot of people are hesitant — but AI is one of the most democratizing technologies we’ve seen. It can also democratize learning.

Keith: Could AI become its own degree, like law or accounting? Sydnee: Maybe, but I think AI is more effective when it layers into other disciplines — product, development, etc. The best applications improve existing workflows. A separate degree might be too narrow.

Keith: Even with all the certifications, I bet your family still asks you for tech help! Sydnee: Every week. My mom especially! “Why won’t ChatGPT do this thing?” It’s actually fun to explain prompts and see them get better results — without needing to understand the underlying tech.

It reminds me of the early days of the internet, when people didn’t even know what a browser was. Keith: Sydnee Mayers, thanks again for joining us on Today in Tech. Great discussion about certifications. Sydnee: Thanks, Keith. It was a great conversation. Keith: That’s it for this week’s episode.

Be sure to like the video, subscribe to the channel, and drop your thoughts in the comments. New episodes every week on Today in Tech. Thanks for watching.