Parents used to tell their children to go into law, banking or the “Big Four” if they wanted to get rich quick, but — if the last 12 months are anything to go by — the next generation of cash-hungry grads will be filing into data science and machine learning.
PhD students are today taking home salaries of up to $400k straight out of uni, while top execs are earning well north of $1m — and fractional execs can pull in even more.
That’s according to four specialist AI recruiters, who’ve witnessed a year of corporate madness since ChatGPT sent the tech world into a spin.
Panic in the boardroom
Many top execs at large corporates had forgotten that AI was a thing when OpenAI released its hugely popular chatbot last November, according to Veena Marr, a digital and tech consultant at executive search firm Spencer Stuart.
“I think panic is the right word, around: ‘Oh my goodness, have we missed the boat? What are we doing? Around the boardroom table we were talking about AI three, four years ago and now we haven't been talking about it for a little while. What does this mean for our competitive advantage? What does this mean for our growth?’” she says.
Many companies are still trying to figure out what kind of talent they need, as they all know they need to be doing something with AI, but they’re just not quite sure what their AI strategy is.
“There is a mix between a lot of excitement and a lot of confusion,” says Andrea Splendiani, EMEA AI recruitment lead for Egon Zehnder, a global executive search firm. “Everybody knows they need to have AI as a top priority. Not everybody knows exactly what that means, especially in terms of talent.”
What kind of AI talent do people want?
What kind of AI talent you need depends a lot on what kind of organisation you are. Splendiani says that larger, less technical enterprises are trying to hire senior AI product leaders — people who’ve worked at AI-native companies and have experience of building with models.
“It’s people who understand the costs and the limits of models. They know the difference between different foundation models and the opportunities that different APIs provide, and how that might change in the future,” he says.
Splendiani adds that most corporates don’t need deep in-house data science expertise, as most enterprise AI products do not require a model to be fine tuned (a technically demanding process where a model is improved using additional data).
These more technically qualified machine learning specialists — with relevant PhDs or experience at top AI labs — are more in demand from tech and AI scaleups, according to Sam Burman, global managing partner of AI talent at executive search firm Heidrick & Struggles.
“They're the ones who are going to be more proactive in the recruitment of that talent coming out of university with a PhD or a master's in AI, because they can plug that talent into their ecosystem far easier,” he tells Sifted.
The recruiters who spoke to Sifted agreed that there hasn’t yet been a huge proliferation of demand for new AI roles like “prompt engineer” or chief AI officer, but Niall Wharton, associate director of data recruitment firm Xcede, says demand is increasing for people who are skilled in maintaining models once they are in place.
“Companies need an experienced LLM professional who can create products from their models, but most importantly, maintain them,” he says. “The titles that have come out of that as being really key have been ML ops engineers… That’s created new roles.”
Let’s talk about some silly AI salaries
So how much are all of these very scarce, very in-demand specialists getting paid in today’s AI gold rush?
Burman says that it’s not uncommon for PhD graduates from top schools in the US to be getting $400k salaries straight out of uni, while the equivalent on this side of the Atlantic ranges from $100k-200k.
In Europe, he says that the top universities for recruiting AI talent are Oxford, Cambridge, Imperial and UCL in the UK, EPFL and ETH in Switzerland, TUM in Germany and the universities of Ecole Polytechnique and ENS in France.
One investor with experience of working closely with people who’ve worked at DeepMind, who asked to remain anonymous, tells Sifted that people with experience at top AI labs are routinely now being offered $800k salaries by recruiters.
When it comes to a more senior level, the highest earning data, analytics and AI executives are earning $2.6m annually in the states, compared to $1.25m in Europe, according to a 2023 compensation survey conducted by Heidrick & Struggles.
Guns for hire
And that’s not even the most you can earn today as an AI specialist, as many are opting to work on a fractional, project-based basis, to draw even more money.
“We are increasingly seeing [AI specialists] make a decision to move into more of an interim career versus a sort of a nine-to-five, permanent employment. It gives them freedom to choose the AI project they want to get involved in and, candidly, the opportunity to earn more money,” says Burman.
“These projects are normally three months at least, but then you get sort of ‘SWAT team’ projects where you've got a real urgent requirement, and it's more like a six-week engagement.”
He adds that the rates for these urgent, time-sensitive projects are basically an “open cheque book” due to the scarcity of talent in the market.
All of the recruiters that Sifted spoke to for this piece say that, along with the companies they serve, recruiters are having to constantly adapt as the boom in AI creates new dynamics in the talent market. But with salaries like these flying around, and the commission to boot, no one in the industry is complaining.
