Picking this week’s jobs to feature was a struggle - too many interesting roles, too little newsletter. Read through what these companies actually need their marketing leaders to solve, and you realize just how far the profession has wandered from Cannes.

Modular robot factories need credibility. Nuclear reactors need neighbors not to panic. Driverless cars need passengers to forget about the missing driver. Fertility clinics need someone to answer the phone. And a gamified university built for goldfish attention spans needs employers to give a shit about its graduates.

Before we get into this week’s picks, a quick look at the wider U.S. market for in-house marketers.

There are 36,618 live jobs this week, up 9.1% year-on-year. Of those, 5,158 are Director level and above, up 16.4%, so senior hiring is outpacing the market overall. 55% of postings disclose salary information, with a median advertised salary of $95,439 across all levels. Here’s how the salaries break down.

Now, onto the jobs, and the rather different problems that come with them.

Foundry Robotics wants to build a factory that can make almost anything.

Founded in 2025 by robotics engineer Adarsh Kulkarni, Foundry is building what it calls the Everything Factory: an AI-first contract manufacturer whose robotic cells can switch between complex products without rebuilding the production line each time.

Foundry’s pitch is not that robots are cheaper than people. It is that adaptable robotics could reduce the supplier density, tooling expertise and production infrastructure required to manufacture complex products in the United States.

Kulkarni puts it this way: “the factory itself is the robot.”

The idea is simple to describe and difficult to prove: make low-volume, complex manufacturing more like software than conventional factory engineering.

Kulkarni has spent most of his career heading toward this problem: robotics at Penn’s GRASP Lab, autonomy at Ghost Robotics, machine-learning infrastructure for autonomous trucks, then Head of Robotics at Scale AI. Earlier this year, he argued that robotics is improving so quickly that the harder problem is no longer the demo. It is scaling reliable systems through messy real-world conditions and edge cases.

Investors are buying the argument. Foundry has announced more than $19 million in funding, backed by Khosla Ventures, Red Glass, Hanabi, Zero Shot and others.

The team is small but revealing. Tesla is its largest source of incoming talent, alongside robotics company Dexterity and people from AWS, Plaid, LinkedIn and McKinsey.

Go-to-market is being built. Vik Sreedhar joined this year as Head of Commercial GTM after senior sales roles in industrial controls and electronic components.

So the marketer arrives after the technical thesis and commercial motion are already taking shape. Their job is to make both believable to people who know exactly how difficult this is.

The JD wants you to explain Foundry to AI researchers, manufacturing engineers, defense program leads and investors. You will own positioning, product marketing, the website, social, technical content, events, CRM, pipeline measurement and agency partners.

There’s always a wrinkle, and it’s usually wearing orange makeup - Washington wants American manufacturing back.  Trump’s policies are also making American factories more expensive to build.

The Trump administration now applies tariffs as high as 50% on many imported steel, aluminum and copper products, while construction inputs for new non-residential projects were 8.9% more expensive in August than a year earlier.

Then there is voltage. Foundry is currently based in California, where industrial electricity averaged 25.11 cents per kWh in July, versus 9.77 cents for the U.S. average.

Foundry does not have to build future factories in Emeryville. But once you start talking about rebuilding American manufacturing, site selection, power and construction costs come to the foreground.

China’s manufacturing advantage is not simply cheap labor, cheap factories or cheaper components. It is the ecosystem. A factory in Shenzhen or Dongguan sits inside a dense network of component makers, tooling specialists, mold shops, electronics suppliers, contract manufacturers, logistics providers and engineers who have spent decades solving production problems together.

Need a component changed? The supplier may be nearby. Need a new fixture? Somebody has probably built something similar before. Need to ramp volume? The skills, suppliers and production capacity are already clustered around you.

I just got back from Shenzhen and I promise you – density is difficult to recreate.

This month, U.S. battery startup EnerVenue opened its first factory in China after abandoning a planned Kentucky facility. Its CEO cited China’s supplier ecosystem, engineering talent and lower costs.

That is closer to Foundry’s real competitive problem than another robotics startup. America cannot recreate decades of supplier density quickly. Foundry’s thesis is that adaptable robotics could reduce how much of that ecosystem has to be created in the first place.

If one production system can be reconfigured through software rather than rebuilt around each new product, the economics of lower-volume domestic manufacturing start to change. That is really the claim this marketer has to make credible.

If Taligence had this search, we’d spend time around Tesla, Anduril, Shield AI, Hadrian, Dexterity, Figure, Palantir and Scale AI: marketers used to technical founders, skeptical engineers, government buyers and products where credibility matters more than polish.

Foundry Robotics doesn’t need to sound bigger, or more patriotic - it needs to sound true.

Foundry is trying to rebuild manufacturing flexibility; our next job clocks in at Antares, which is trying to put factory-built nuclear power into military bases and eventually space. The company has also crossed an important line from pitch to proof: its Mark-0 achieved zero-power criticality at Idaho National Laboratory on June 4, making it the first private advanced reactor to do so under the DOE pilot program.

Since then, it has raised $470M Series C, with the capital aimed at moving from demonstration toward deployment, including U.S. military installations.

That makes the marketing hire beautifully awkward. The product is now real enough that communications cannot be pure muskism - but also cannot assume trust.

And the JD is extraordinary: Marketing sits in Government Affairs, reports to the Chief Nuclear Officer and must influence military services, DOE, NRC, Congress, state governments, Tribal governments, host communities, investors and customers while owning crisis preparedness.

OK.

The company has enough proof that the story no longer needs to rely entirely on PowerPoint. Now the job gets serious: deployment, regulation, fuel, manufacturing, customers and public acceptance.

That last one is probably the hardest.

People hear “microreactor” and nuclear engineers may think compact, inherently safe power generation. Plenty of normal humans just hear the second half of the word.

That's why I'd be wary of a conventional tech marketer. Poor OpenAI have already failed to hire a Chief Communications Officer in the last 9 months. This job means explaining complex technology without terrifying or patronizing people.

The strongest candidate may not be called a marketer by title today. They need Washington instincts, industrial credibility and enough technical curiosity to sit with nuclear engineers without sloganeering.

Antares wants someone who has taken a company through a "defining public moment." That tells you where management thinks it's heading.

Antares’ reactor sustained a controlled nuclear chain reaction. Impressive - and still the easiest part of the communications problem.

Last week I took a Pony.ai taxi in Shenzhen. Nobody was driving it. We giggled when we got in. The strangest part was how quickly it stopped being strange. That is more or less the marketing problem facing Waymo, and they know that trust is built through experience.

The company is hiring a Senior Global Experiential Marketing Manager to build sponsorships, events and physical experiences around autonomous driving. Normally I would struggle to get excited about an events job, but this one is different.

Waymo says the marketing team is responsible for building “societal acceptance for autonomous driving technology for the first time in history.” This hire is specifically expected to turn physical experiences into public trust.

That sounds grand until you remember what they are selling: A car turns up. You open the door. Nobody is sitting in the front. Then you let it drive you into traffic. No amount of performance marketing removes that first moment of hesitation.

It has now completed more than 20 million fully autonomous rides and is operating across 15 major U.S. cities, with expansion underway internationally. Earlier this month it announced Singapore for 2028; Tokyo is planned for 2027.

Waymo's latest analysis covers 270 million fully autonomous miles across 5 major U.S. markets. Against its human-driver benchmark, the company reports 82% fewer injury-causing crashes and 95% fewer crashes involving serious injury or worse. Those are Waymo's numbers and methodology, but the sample is now large enough to matter.

So autonomous driving has moved into a different phase. For years, the category was sold mostly through bullshit.

Like these giant fibs from Elon Musk, which have tainted the category. In April 2019, he said Tesla would have more than one million robotaxis on the road in 2020. It did not. Tesla also stopped allowing customers to buy their leased cars at the end of their contracts because Musk said those vehicles would be needed for the coming robotaxi fleet. Many were eventually refurbished and sold instead.

Seven years later, Tesla finally has a real autonomous ride-hailing operation and has begun limited public Cybercab rides in Austin. But scale is still rather different from the old forecasts. Reuters reported this month that Tesla had 420 autonomous vehicles registered in Texas, including 45 Cybercabs.

Meanwhile, I can get into a Pony.ai car in Shenzhen. Pony had 1,975 robotaxis operating by the end of Q2, with services in China's largest cities and Robotaxi revenue up 691% year-on-year. Its Shenzhen network now reaches Bao'an Airport, Shenzhen Bay Port and Shekou Cruise Port.

That is why I think Waymo's real job is being the first to make normal people comfortable using it. Waymo's JD is unusually honest about that. The hire will own sponsorships, partnerships, demonstrations and events, with one objective appearing repeatedly: trust.

And experiential marketing makes unusual sense here. You can tell someone a robotaxi has driven 270 million autonomous miles. You can show them safety charts. You can explain LIDAR.

Nothing is quite as persuasive as sitting in the back seat while the steering wheel turns itself through an intersection and discovering that, actually, you are mostly checking your phone.

The marketing job is therefore less about making autonomous driving look futuristic and more about making it look boring, safe, predictable, ordinary.

The wrong marketer will lean into spectacle: giant CES demonstrations, sci-fi language and people filming an empty driver's seat for TikTok. The right one will understand that Waymo wins when the empty driver's seat stops being the interesting part of the journey.

The candidate pool should reflect that.

If Taligence had this search, I'd look beyond conventional automotive experiential marketers. Apple Retail, Disney, Live Nation and major sports sponsorship teams all understand how physical experiences can change feelings, so that’s where I’d look.

Nobody is buying a shoe or a plastic model of the car at the end of the activation. They are deciding whether they trust a machine with their body.

Waymo has already built the car, driven the miles and accumulated the safety data. So this job is simply: Get people into the back seat.

I don’t have children. I don’t particularly like children (although my nephews are aces), but I do appreciate a job description that tells a CMO exactly how they will be judged.

Ivy Fertility is hiring its first Chief Marketing Officer. The company now has around 70 physicians across 30 locations in 9 states, built partly through an aggressive acquisition strategy backed by private equity firm InTandem Capital.

And the org chart tells us almost everything. Constance Rapson, who spent nearly four years as Chief Growth Officer, has just moved into a new role as Chief Brand Officer. Her remit now includes how Ivy is “understood, trusted and valued,” international expansion and brand partnerships. The new CMO gets patient acquisition.

There’s the confession. Brand stays with the Chief Brand Officer. The CMO gets the spreadsheet.

The JD is pretty damn specific: increase consult volume, improve lead-to-consult conversion, lower cost per scheduled consultation, lower cost per new patient, grow marketing-sourced revenue and manage acquisition by clinic and service line.

They even want Marketing working with operations so demand matches available appointment capacity by location.

Ivy is not one giant fertility clinic. It is a network of local practices, many with names and reputations built long before Ivy existed. The group has been assembling them since 2022 and has announced around 10 acquisitions, most recently Santa Barbara Fertility Center this month. So you can’t just buy more Google traffic.

Generate too much demand in the wrong city, and there may not be enough appointments. Push the wrong service line and the economics change. Centralize branding too hard, and you torch the local reputation you paid to acquire.

Employer-funded fertility has become a much bigger category. And patients don't shop for IVF like sneakers. They arrive anxious, comparing success rates, cost, wait times…bombarded by identical promises.

Ivy's network makes it harder. A national platform wants scale. Reputation lives at clinic level, or with the practitioner. Hence the KPI: 90% call answer rate. A beautiful campaign is worthless if a frightened patient calls and nobody answers.

This is a PE-backed roll-up. Monthly Board reporting on CPA, funnel conversion, ROI. No hiding behind brand health. The CMO doesn't need to make more people want children. They need to make sure that when people decide they do, Ivy gets the appointment.

Outsmart wants to reinvent college. Its first senior marketing hire will be measured on Daily Active Users. That should make everyone nervous.

Former Duolingo executives Jorge Mazal, Gina Gotthilf and Daniel Falabella founded the company. Mazal was Duolingo’s Chief Product Officer; Gotthilf ran Growth and Marketing. The thesis is obvious enough: take personalization, habit formation, experimentation and product-led growth, then point the machinery at higher education.

Investors like it. Outsmart has raised $38 million, including a $25 million round backed by DST Global Partners, Forerunner, Khosla Ventures and Lightspeed. The company is still early, hiring across product, engineering, learning science and academic roles while preparing for a broader launch.

The Founding Marketing Lead we feature reports to Gotthilf, who is already co-founder and CMO…Outsmart already has people who know how to manufacture engagement.

The JD wants repeatable growth loops, creators, social, PR, brand and reputation across students, parents and the education market. It says the company is already generating millions of organic views every week.

Then you get to the north star: DAU.

I use Duolingo. I’ve used it for years. It is fun. It is brilliantly designed. It is extremely good at making me open Duolingo again. I learned far less Chinese from it than my streak might suggest. Mostly, I have become good at Duolingo, the game.

I know how to protect the streak, collect points, exploit the exercises and avoid losing imaginary hearts. A questionable credential.

That criticism is hardly novel. Duolingo has spent years turning language learning into a game, and for plenty of users the game eventually becomes the product. You can be extremely engaged but still shit at the language. That is annoying when the stakes are a few minutes on your phone.

It’s a bit much when somebody starts calling the product a university.

I’m a recruiter, so I have seen the other end of this industry: the endless supply of expensive creds that mean far more to the person who bought them than to anyone doing the hiring (hello 99% of “AI” certs)

For experienced candidates, academic credentials barely register most of the time. In volume hiring, the CV gets five seconds: summary, employer, title, first three bullets.

Nobody is lovingly inspecting the certificate from the Institute of Transformational Digital Leadership. But an 18-year-old has a different problem. They do not yet have the employer, title or bullets. For them, education is partly supposed to help create those things.

Outsmart cannot simply build an enjoyable learning product and congratulate itself when DAU goes up. At some point, somebody has to give these people jobs.

Students need to want it. Parents need to believe they are not funding an expensive experiment. Employers need to know what an Outsmart graduate can do. Recruiters need to understand whether the qualification means anything at all. DAUs don't solve any of this.

Did anything valuable happen? If Outsmart gets that wrong, it could produce extraordinary engagement, beautiful retention curves and graduates waving a credential nobody gives a rats’s arse about. (Which would still look fantastic in the dashboard)

Which is why I would push hard on DAU being the queen metric in an interview. Not because it is useless. For an early consumer product, it is a fair measure of whether people like using the thing.

The founding marketer needs to understand the difference between acquisition and legitimacy. Plenty of people can buy attention; plenty more can manufacture engagement.

If I were doing this hiring, I would look at people from Masterclass, Handshake and LinkedIn, but I would be suspicious of anyone who talks too much about community, creators and viral growth without eventually getting to employment outcomes.

Because Outsmart could genuinely build a better route into work than a lot of universities currently provide. It could also build a very expensive game with shitty certificate at the end.

The real test is whether anyone hires users after their DAU’ing.