Posted by Dylan Hoyle – 01/09/2026
Founder-Led Sales Doesn’t End. It Gets Delegated Badly.
Ask a Founder when Founder-led sales ends and most will give you a date – usually the start date of their first AE. Ask that AE six months in, and you’ll hear a different story.
The handover from Founder to first sales hire is one of the most fumbled transitions in early-stage companies. Not because Founders delegate too little, but because they delegate the wrong things, at the wrong moment, with the wrong expectations.
Founder-led sales doesn’t end. It changes shape. The companies that get this right treat the first AE as an extension of the Founder’s selling motion, not a replacement for it.
Posted by Dylan Hoyle – 01/09/2026
Territory, Quota and Comp Design When the Market Doesn’t Exist Yet
Most quota problems at early-stage Infrastructure Companies aren’t performance problems. They’re design problems – targets set by borrowing numbers from markets that behave nothing like this one.
There are no comparables for a category you’re creating. No benchmark tells you what a reasonable first-year number looks like for a product buyers didn’t have a budget line for a year ago.
So Founders guess high, attach an aggressive comp plan to the guess, and watch a good hire miss a number nobody could have hit.
Posted by Dylan Hoyle – 01/09/2026
Technical Credibility Is Now a GTM Requirement
The buyer changed before the seller did. In developer tooling and AI infrastructure, the person who decides whether your product gets adopted is an Engineer – and Engineers have no patience for a seller who can’t hold the technical thread.
That has quietly rewritten the spec for GTM hiring. Technical credibility used to be a nice-to-have for AEs, PMMs and sales leaders. It’s now the difference between a first meeting and a second one.
But most Founders assess it badly – either skipping it entirely, or turning the interview into an engineering exam. Both miss the point.
Posted by Dylan Hoyle – 19/08/2026
European AI Companies Hire Into Two Markets at Once. Most Plan for One.
A European AI company raises a Series A, and the board deck has one hiring plan on it. In practice there are two.
One runs in Paris or Berlin or Tel Aviv, where the research and engineering density is, where the Founders’ network is, and where the interview process already works. The other runs in New York or Austin or the Bay Area, where the buyers are. They have different compensation bands, different notice periods, different definitions of a good first seller, and different failure modes.
Most Companies write the first plan carefully and improvise the second. This is a model for treating them as one.
Posted by John Hitchen – 19/08/2026
The US AI Talent Market Is Four Markets. Most Hiring Plans Assume One.
Most US hiring plans carry one number for a role. One title, one band, one timeline, applied wherever the person happens to live.
That number is an average of four markets that behave nothing like each other. Bay Area Research Talent, New York enterprise Go-To-Market, Seattle Infrastructure Engineering, and a distributed population that left the hubs and is not coming back.
Averaging them produces a plan that is too expensive in one market, uncompetitive in another, and slow everywhere.
Posted by Adam Richardson – 05/08/2026
Machine Learning Engineer Is Four Different Jobs. Most Briefs Only Describe One.
Machine Learning Engineer is the most overloaded title in AI hiring. Four distinct jobs sit underneath it, and most briefs describe none of them precisely.
That is how a Series A company spends three months interviewing research scientists for a role that needed an inference engineer, then concludes the market is short of talent. The market was fine. The brief was not.
Here is what the four profiles actually are, how to work out which one you need, and what to test for once you know.
Posted by Dylan Hoyle – 01/08/2026
Your First Solutions Engineer Decides Whether the POC Closes
In Enterprise AI, the demo does not close the deal. The proof of concept does.
And the person who runs that proof of concept is usually not the account executive. It is the solutions engineer – the hire most Seed and Series A infrastructure companies delay until three deals have already stalled in evaluation.
Here is what the founding solutions engineer actually owns, why the profile is different in AI and data infrastructure, and how to evaluate one when you have no technical sales function to benchmark against.
Posted by Dylan Hoyle – 01/08/2026
Sequencing GTM Hires at Seed and Series A
Most Founders hire in the order roles become painful. Not the order the company needs them.
That is how a fifteen-person developer tooling company ends up with two account executives, no solutions engineer, no product marketer, and a founder still writing every piece of positioning at midnight. Each hire was defensible on the day it was made. The sequence was not.
Here is a sequencing model for go-to-market hiring at Seed and Series A, the triggers that should move each role forward or back, and the two mistakes that cost the most.
Posted by Dylan Hoyle – 01/08/2026
Hiring GTM Talent Out of Big Tech Into a 20-Person AI Company
The strongest CV in an early-stage GTM search is often the weakest signal.
A seller with eight years at a hyperscaler and a run of quota-beating years is easy to say yes to. They know Enterprise. They know the Buyer. They have closed numbers larger than the company’s entire ARR. And roughly half of them will be gone within a year of joining a twenty-person AI company.
The failure is rarely about capability. It is about what the number was actually built on, and how much of that came with them.
Posted by Dylan Hoyle – 21/07/2026
Hiring at an AI Startup Is a Judgement Problem, Not a Sourcing One
Most Founders think their hiring problem is a pipeline problem. Not enough good candidates, not enough time, not enough reach.
At an AI startup, that is rarely the real issue. The best engineers and operators are known, and already being courted by three other companies. Finding them is not the hard part.
The hard part is judgement. And judgement is what most searches get wrong.
Posted by Dylan Hoyle – 20/07/2026
What Enterprise AE Compensation Actually Looks Like in AI Infrastructure
Compensation is where early-stage AI infrastructure companies quietly lose their strongest GTM candidates. Rarely on the base. Almost always on the structure.
A Founder benchmarks against a general SaaS survey, offers a 50/50 split against a quota nobody can hit, and watches three strong candidates decline inside a fortnight. The offers were not mean. They were designed for a market with comparables, in a market that has none.
Here is what Enterprise AE compensation actually looks like at Seed and Series A in AI and data infrastructure – and where the design decisions matter far more than the headline number.
Posted by Dylan Hoyle – 20/07/2026
What Candidates Ask AI Startups That Founders Aren’t Ready For
Five years ago a senior candidate joining an early-stage company asked about the product, the team and the funding round. Today they ask about gross margin under inference cost, what happens to the product if the next frontier model absorbs the feature, and whether the data rights in the enterprise contracts actually hold.
Most Founders are not ready for those questions. The unprepared answer is not usually a bad answer – it is a defensive one, and candidates read defensiveness as risk.
The interview now runs in both directions. Here are the questions strong candidates are asking, and what a credible answer sounds like.
Posted by Dylan Hoyle – 20/07/2026
When to Hire a Head of Sales – And Why Most AI Startups Do It Too Early
The number one mistake I see Founders make with their first sales leadership hire is treating it as a rescue.
The role gets opened after a hard quarter – pipeline thin, Founder stretched, board asking how revenue scales – and a Head of Sales looks like the answer.
Nine months later the company has spent a quarter of a million pounds, cycled two AEs, and still has no repeatable motion. The leader leaves. The Founder goes back to selling.
The trigger for this hire is not a revenue number, and it is not exhaustion. It is evidence that the same sale can be run twice by someone who is not you.
Posted by Dylan Hoyle – 20/05/2026
How to Run an Interview Process at a Startup
Most startups lose great candidates not because of competition – but because of their own process.
Long gaps between stages. Interviewers who haven’t been briefed. Offer conversations that start two weeks after the final round.
At the early stage, your interview process is your employer brand. Here’s how to run one that actually works.
Posted by Dylan Hoyle – 20/05/2026
Recruiting Your Founding PMM in AI / Infra
Most AI infrastructure companies wait too long to make this hire.
They’ve got strong engineering. Maybe an early sales leader. A founder who’s been carrying the narrative. And then they hit a wall – the messaging isn’t landing with the right audience, sales doesn’t have the materials they need, and nobody has time to fix either.
Here’s what you need to know about making this hire well.
Posted by Dylan Hoyle – 28/04/2026
The War for Talent in AI
Everyone is hiring the same ten people.
That is not an exaggeration.
In AI & infrastructure, the pool of candidates who combine technical depth, commercial credibility, and genuine community presence is extraordinarily small.
And every well-funded startup in the category knows exactly who they are.
Posted by Dylan Hoyle – 07/04/2026
The DevRel Hire That Actually Moves the Needle
Developer relations is one of the most misunderstood roles in AI infrastructure.
Most companies treat it as content.
A blog post. A YouTube tutorial. A conference talk.
That is not what the role is.
Done correctly, DevRel is a distribution channel, a product feedback loop, and a trust-building engine – running simultaneously.
Posted by Dylan Hoyle – 25/02/2026
The First Product or Engineering Hire in an AI Infrastructure Startup
In AI infrastructure companies, product and engineering hires are not incremental additions.
They define:
Architecture
Velocity
Long-term defensibility.
And often shape the technical DNA of the company.
