Vector

Most companies file DevRel under marketing. That is why the hire so rarely works. A blog post. A conference talk. A YouTube tutorial. Treated as content, the role produces content....and very little else. Done properly, it is three things at once: a distribution channel, a product feedback loop, and a trust engine with the developers who decide whether the product gets adopted. In AI infrastructure, that combination is closer to go-to-market than it is to comms. Read the full piece here: https://lnkd.in/g366Bdqk #withvector

Most Frontier AI Founders hire one ML profile when the roadmap needs two. The job spec says "ML engineer." The instinct is to find someone brilliant and let them cover everything. But research and production are different disciplines, not two ends of one spectrum. A research scientist pushes the frontier of what a model can do. An ML engineer makes it run reliably in front of real users. Where the conflation quietly costs time: • A researcher hired to ship infrastructure, frustrated within months • An engineer hired to invent methods, stuck without a research culture • A brilliant generalist stretched across both until they do neither well • A roadmap that assumes one hire closes a gap that needs two The sharper move is naming which problem comes first - frontier capability or production reliability - and hiring directly to it. Hire one profile for both, stall on both. Hire to the real problem, move on both. Which of the two does an early Frontier AI team usually need first? #withvector

The strongest CV in an AI infrastructure search is often the weakest signal. The logo does a lot of quiet persuading. A Frontier lab on the CV. A FAANG team. A name everyone recognises. The instinct is to read that as proof. But a brilliant operator inside a 2,000-person org solved a very different problem than a Seed-stage company is hiring for. At that scale, the systems already exist. The roadmap is set. The budget is approved. Zero-to-one is a different sport. Where the logo quietly misleads: • Pedigree that signals execution inside structure, not building it from nothing • Comfort with resources a Seed company will not have for years • A track record of optimising, not inventing • Specialisation so deep it never had to flex across functions The sharper signal is rarely the brand on the CV. It is whether the person has built something before the playbook, the team, or the budget existed. Hire the logo, inherit an operator waiting for structure. Hire the builder, get someone who creates it. What is the clearest sign a candidate can build from zero, not just operate at scale? #withvector

𝗪𝗲'𝗿𝗲 𝗵𝗶𝗿𝗶𝗻𝗴 𝗮𝘁 𝗩𝗲𝗰𝘁𝗼𝗿! Vector is a specialist firm. We hire for AI infrastructure, data/cloud, and developer tooling. We don't do everything - that's the point. If you're a recruiter who's tired of being average at a generalist firm, or someone early in your career who wants to start in the right environment - we'd love to talk. Ideally Tampa based, but also open to recruiters interested in relocating to Sunny Florida. Reach out to Dylan Hoyle for more information or apply here: https://lnkd.in/eez5tkgJ #hiring #recruitment #tampajobs #ai

Excited to share that our Founder Dylan Hoyle, was recently on the THE SEARCH PODCAST discussing scaling a start-up and making the move to the USA. Links below, check it out! #withvector #podcast

In a talent pool this small, the reference call is the real interview. The formal process gets the attention. Four rounds, a take-home, a panel debrief. Then the decision quietly turns on a two-minute backchannel call to someone who worked with the candidate three years ago. In AI and infrastructure, the people who can do these roles all know each other. Signal travels fast. Most teams still run references as a formality: • Calling only the names the candidate provided • Asking questions that invite a polite yes • Treating it as a final box-tick, not a source of truth • Skipping the backchannel because it feels awkward The sharper teams treat references as primary evidence, not confirmation. They map the candidate's real network, not the curated list, and ask what the person is genuinely great at...and what they are not. Curated references confirm a decision already made. Backchannel references change it. Where has a reference call actually changed a hiring decision this year? #withvector

Most startups don't lose great candidates to competitors. They lose them to their own process. Long gaps between stages. Interviewers walking in cold. Offers that land two weeks after the final round. At the early stage, the interview process is the employer brand. Every delay and every unbriefed conversation tells a candidate what the company is really like to work for. The teams that win talent treat their process as a product - fast, deliberate, and respectful of the person on the other side. Read the full piece here: https://lnkd.in/eKFT_K9E #withvector

The most important hire in an agentic systems company often has no title yet: the Forward Deployed Engineer. Most Founders file it under support, or assume a strong AE will cover the gap. Neither holds once the product meets a real customer environment. Agentic systems do not demo and deploy the same way. The distance between a clean demo and a production workflow is where deals quietly stall. The FDE lives in that distance. What the role actually carries: • Translating a customer's messy workflow into something the product can run • Earning trust with the engineers who will decide whether it gets adopted • Feeding production reality back into the roadmap, every week • Turning a single deployment into a repeatable implementation pattern Founders often hire this person too late....after the first enterprise deal has already slipped. The sharper move is hiring the bridge before the gap becomes the reason a deal dies. Hire it as support, get tickets closed. Hire it as engineering at the front line, get adoption that sticks. Where should the first forward-deployed engineer sit? Product, Engineering, or GTM? #withvector

Most hiring decisions in AI infrastructure still come down to a feeling in the room. Strong interview. Good rapport. Everyone nods. An offer goes out. Three months later the hire is struggling, and nobody can say exactly what was missed. The problem is rarely the candidate. It is the absence of a structured evaluation framework. Gut feel rewards the person who interviews well, not the one who does the job well. A structured framework changes what actually gets measured: • Defined competencies agreed before the first interview • A consistent scorecard every interviewer completes independently • Evidence captured against each signal, not a vague summary impression • A debrief that compares notes rather than averaging vibes The point is not to remove judgement. It is to make judgement comparable across the panel. Unstructured interviews measure charisma. Structured ones measure the work. Which competency is hardest to assess in an interview for a technical GTM or engineering role? #withvector

𝗪𝗲'𝗿𝗲 𝗵𝗶𝗿𝗶𝗻𝗴 𝗮𝘁 𝗩𝗲𝗰𝘁𝗼𝗿! Vector is a specialist firm. We hire for AI infrastructure, data/cloud, and developer tooling. We don't do everything - that's the point. If you're a recruiter who's tired of being average at a generalist firm, or someone early in your career who wants to start in the right environment - we'd love to talk. Ideally Tampa based, but also open to recruiters interested in relocating to Sunny Florida. Reach out to Dylan Hoyle for more information or apply here: https://lnkd.in/eez5tkgJ #hiring #recruitment #tampajobs #ai