Snorkel AI's $3.5 billion valuation and the AI jobs around the model

Snorkel AI just raised $350 million at a $3.5 billion valuation, with its CEO pointing to surging demand for complex AI training data.

The less glamorous problem: teaching the model

For the last few years, everyone has been obsessed with models.

Who has the biggest model? Who has the best benchmark? Who has the most GPUs? Who can train the next generation of AI?

But eventually you run into a much less glamorous problem: your model needs to learn something difficult.

And the data required to teach it isn't sitting neatly in a database waiting to be downloaded. How do you even begin to teach heuristics to a computer?

You might need an expert to define what a good answer looks like. You might need engineers to build the environment where an AI agent can be tested. You might need researchers to figure out why the model keeps failing at a particular class of problems.

That's exactly the kind of work Snorkel is building around. The company describes its business as developing specialized training data, evaluation environments and solutions for frontier AI systems.

The roles appearing around that work

And look at the roles appearing around that work. We even landed a Head of Data Quality interview at Snorkel for one of our clients back in July.

Snorkel's current openings include senior AI and platform engineers, security engineers, forward-deployed engineers, product managers for data and infrastructure, research scientists focused on evaluation and training, and technical project managers for enterprise AI.

Finally, the AI job market has started to expand sideways.

You don't necessarily need to be the person inventing the next foundation model. There is a growing amount of work around the model: data, evaluation, infrastructure, security, deployment, agent environments, domain expertise and productization.

What did you actually own?

For senior technology leaders, "I have worked on AI" is becoming a rookie statement.

What did you actually own?

Did you build the platform that supported it? Did you solve a data-quality problem that was preventing the model from working? Did you put AI into production? Did you build evaluation systems? Did you make an agent reliable enough for a real business process?

Those are much easier things for a hiring manager to understand, and this is going to matter even more as companies become more disciplined about AI spending.

The easy projects will get cut. The projects that require serious engineering and produce measurable business value will keep getting funded.

So if you're trying to position yourself for the AI job market, don't just chase AI keywords.

Find the difficult layer underneath the AI and get good at it.

If you’re looking for such roles but can’t land interviews, we have you covered =)