Head of Marketing

This is position will support US operations and will work US Shift, Please only apply if you have current/past experience working US shift and 100% committed to working US Shift

# Head of Marketing

We built the marketing org before we hired the marketing leader. Now we are hiring the leader.

Curately is an AI recruiting platform with 40+ staffing customers and 9 enterprise workforce programs, and until this summer it had no marketing team at all. Since July the founder has built one out of software: seven AI marketing agents, the governance and knowledge base they run on, and the automation that runs them on a schedule. It produces real output every week, and it is being documented in public, failures included.

You will take over that org and own everything marketing is accountable for: top of funnel, pipeline, and ultimately revenue. The title can be Head of Marketing or VP of Marketing, depending on your level. The job is fixed: you are the general manager of marketing, its product marketer, its growth operator, and its GTM engineer, leading a team that is mostly agents plus a small group of human specialists you will grow deliberately.

## About Curately

Curately is an AI Candidate Conversion Platform for staffing firms and enterprise talent teams. It covers the top of the recruiting funnel end to end, across three pillars:

- **FIND**: AI Sourcing in plain English across 200M+ sourced passive profiles (US) and the job boards recruiters already use; AI Match, which surfaces the best-fit candidates already sitting in the ATS; and a Chrome extension that works on top of the ATS, so recruiters never switch tabs.

- **ENGAGE**: Maya, one multi-modal AI agent that reaches candidates by voice, email, and SMS, around the clock.

- **CONVERT**: qualification, routing, scheduling, and Journeys that carry a candidate from first contact to a filled role.

From find to filled.

Who buys it: 40+ staffing firms and 9 enterprise contingent-workforce and direct-sourcing programs. Curately integrates with the ATSs staffing firms actually run (JobDiva, Bullhorn, Avionté, Ceipal, LaborEdge, Tracker, Brightmove, Workable) and sells to delivery and recruiting leadership at staffing firms and to contingent-workforce program leaders at large enterprises. Two further motions, enterprise talent acquisition and self-serve, are defined and not yet proven. Part of this job is proving them, or deciding not to.

## What you will inherit

Most marketing leaders join a company and inherit an empty org chart. You inherit a working system, built in public since July, and the honest state of it.

**Seven AI agents, each with a written charter, a state doc, and a run cadence:**

- **VoC Miner**: turns every customer and prospect call (via Fireflies) into verbatim, verified voice-of-customer cards; hundreds of cards distilled into canonical pain patterns and a language bank that every other agent draws from. Runs weekly.

- **Competitive Intel**: evidence-tiered battlecards, displacement plays, and same-day deal briefs. What customers say about a competitor outranks what the competitor says about itself.

- **Positioning Keeper**: owns the canon: the positioning master, ICP and personas, and the Claims Register and Terminology Registry that govern every public word.

- **Content/AEO Engine**: writes for answer engines, not just search; runs a tracked-prompt portfolio in Peec AI to measure share of voice in AI answers against tracked competitors; ships through the Lovable-built site with a human on the publish button.

- **LinkedIn Advocacy Engine**: a role-based content calendar and per-person voice profiles for the founder and team. The agent drafts; humans post. A weekly performance loop feeds what works back in.

- **Creative Studio**: a shared service that turns approved words into on-brand statics, carousels, ad creative, and video (Kie for generation, Clueso for real product capture), under anti-slop image guidelines and a strict authenticity rule.

- **Demand Gen**: a multi-channel paid brain: channel-agnostic principles, one playbook per channel, one technical adapter per channel. LinkedIn is ratified and live, Meta is fully specified, Google is next. It proposes; a human executes.

**The law they run on.** A governance stack every agent inherits: Operating Judgment (how to apply rules with proportionality and a bias to ship), the Claims Register and Terminology Registry, the Company Voice Guide, a customer-permissions ledger, and evidence tiers (customer-voiced beats firsthand beats public beats relayed). Nothing publishes autonomously. Agents propose; humans approve.

**The knowledge base and the runner.** Canon and run records live in ClickUp: Docs, a Decisions list, an Action Register, and a Run Log. An Agent Org Runner app (Python, on Replit) fetches calls from Fireflies on a schedule, triages them in code, extracts with verbatim quote verification, publishes run docs, opens decisions in ClickUp, and sends a weekday digest, with cadence checks and a kill switch. Behind all of it sits a blueprint of 40+ agent workflows across 11 marketing functions; seven are live.

**The stack.** Claude (the agents reasoning layer), ClickUp, Fireflies, HubSpot, Peec AI, Lovable, Kie, Clueso, Fibbler (LinkedIn attribution), Zapier, Replit, and Google Drive as a frozen archive and brand-asset store.

**Live channels and real data.** A LinkedIn ABM program aimed at the customer base of one of the ATSs we integrate with, with real spend and real, unflattering early numbers. Organic LinkedIn that has been outperforming paid on the same accounts. A Building in Public series on the founder's LinkedIn that documents the whole system, which means you also inherit an audience.

**The humans.** A founder with deep product, ICP, and customer knowledge who has been the org's editor-in-chief and its message bus, and will hand you both seats. Existing SEO and performance-marketing specialists. A sales team, CS and implementation leads, and engineering. Six to ten industry events a year.

**The honest state.** It works and it is unfinished. The founder ran every agent by hand for two months before automating the loop; some cadences slipped, some rulings never reached the master docs, and the first paid campaign went live before its targeting was refined. Every one of those failures is written down, with the fix. You will not be handed a finished machine. You will be handed a real one.

## What you will own

You own marketing end to end, and the number it is accountable for.

- **Pipeline and revenue.** Qualified demos, marketing-sourced and marketing-influenced pipeline, and the revenue that follows, across the proven segments (staffing delivery and recruiting leadership; enterprise contingent-workforce programs) and the two not yet proven (enterprise talent acquisition; self-serve).

- **The AI-agent org.** You are the editor-in-chief. You run the weekly loop, take over the approval queue, ratify claims, and decide what gets built next from the blueprint (customer proof engine, account intelligence, analytics and reporting, launch engine, lifecycle) and what gets retired. The standing rule: an agent that does not move a metric within 60 to 90 days is killed.

- **GTM engineering and RevOps.** HubSpot as the source of truth; funnel definitions and attribution (Fibbler today, a warehouse when volume justifies it); conversion tracking on every channel (Insight Tag, Pixel, server-side conversions); enrichment and real-time lead scoring so channels optimize toward qualified demos rather than demo count; and the Runner's roadmap toward earned, narrow autonomy.

- **Product marketing.** Positioning, messaging, claims discipline, ICP and personas, launches, sales enablement, and competitive strategy. You keep the canon true, current, and used.

- **Demand generation across channels.** Paid social (LinkedIn live; Meta and Google next), organic LinkedIn and founder-led advocacy, AEO and SEO, ABM into named ATS-customer and enterprise accounts, events, customer marketing and proof, and lifecycle.

- **Brand and creative.** Taste is the gate. You decide what looks and sounds like Curately, and you enforce the authenticity rule: no AI imagery presented as real people, real events, or real product.

- **The team.** Lead the existing specialists, hire humans only where they beat agents on judgment and craft, and own the budget for both.

## Who you are

Four things at once. Most candidates have two. We need all four at a working level, and two of them exceptional.

**A general manager's mind with an engineer's habits.** You have run something with a number attached and been judged on it. You think in systems: inputs, throughput, feedback loops, failure modes. You would rather read a run log than a status update, and you fix the process before you add a person.

**A product marketer's brain.** You have owned positioning at a B2B software company and can tell strong copy from plausible copy in ten seconds. You understand that a claim is a liability until it is substantiated. You do not have to be the best writer in the room, but you have to be the best judge of writing in the room.

**A growth operator's record.** You have carried a pipeline number in B2B, mid-market or enterprise, with a real sales cycle. You have run paid, ABM, lifecycle, and events, and you know the difference between a cold-start problem and a targeting problem. You are fluent in a CRM and comfortable in a spreadsheet; SQL is a plus.

**An AI-native builder.** You use Claude or comparable models daily for real work. You have built automations or agents yourself (Claude Code, Zapier, n8n, Clay, or plain code) and shipped something that ran without you. You read a prompt the way an engineer reads code and a schema the way a PM reads a spec. APIs, MCP servers, webhooks, and warehouses do not slow you down. Your instinct is to extend the system, not to ask for headcount.

**Experience.** Typically 8+ years in B2B software marketing, including a stretch as the first or second marketing hire at an early-stage company and time leading a function. Recruiting, staffing, or HR tech is a plus, not a requirement; the founder and the customer-voice base cover the industry.

**Character.** Bias to ship. Proportionate judgment: you know a blocker from a note. You cite evidence and dates rather than opinions, and you change your mind when the data does. You are comfortable being a public voice for the company. You want leverage, not a large team.

## Your first 90 days

**Days 1 to 30: take the seat.** Read the canon: Operating Judgment, the Claims Register, the Voice Guide, ICP and personas, the voice-of-customer master. Run each of the seven agents yourself at least once. Sit in on customer and prospect calls. Take the weekly loop and the approval queue off the founder's desk. Rule on the open decisions the data has already raised: the deal-size assumptions behind our paid thresholds, whether founder-led advocacy should lead paid on the accounts that are actually converting, and the fix-or-pause call on the live LinkedIn campaign.

**Days 31 to 60: make the numbers trustworthy.** Funnel stage definitions in HubSpot, conversion events firing on every live channel, and attribution you would defend to a board. Define what a qualified demo is and stand up lead scoring against it. Relaunch paid with the right audience, or hold it until it can be measured. Ship the advocacy seam (founder and exec posts, amplified) if that is where the pipeline is.

**Days 61 to 90: own the plan.** Bring a growth plan with a pipeline target for the next two quarters and the budget to hit it. Decide which blueprint agents to build next and which live ones to retire. Make your first human hire, if any. Take a first, narrow slice of agent autonomy live (an agent pausing a clearly dead ad, for example) with the guardrails that make it safe.

## How we work

- **Autonomous where it helps. Human where it matters.** Agents draft, propose, and monitor. Humans approve anything that reaches a customer, a prospect, or an ad platform. Spend authority never sits with an agent.

- **Evidence over opinion.** Customer-voiced evidence outranks internal opinion, which outranks public sources, which outranks anything relayed. Every ad angle, page, and post cites what it rides on, with a date.

- **The law prevents harm, not work.** Claims and terminology are governed, and so is the way the rules are applied: the founder is the authority, weigh real risk, flag once then defer, blocker is a high bar, bias to ship.

- **Authenticity.** Real product is real product. Real events are real photos. Generated art is fine when it reads as illustration and is labeled as such.

- **Give, then ask.** Practitioner to practitioner. Positive about every competitor, tool, and ATS. No engagement bait, no filler, no AI tells.

- **Building in public.** How this org was built is being documented on LinkedIn as it happens, failures included. You will be part of that story, in your own voice.

## Is this you

This is you if:

- Seven agents, a knowledge base, and a runner sound better to you than seven headcount.

- You have opinions about where our blueprint is wrong and will say so in the first conversation.

- You can critique our homepage's positioning in twenty minutes and enjoy doing it.

- You have a story about something you built that ran without you.

- You want to be judged on pipeline and revenue, not on activity.

This is not you if:

- You expect to hire a team before you ship anything.

- Your first instinct with an AI draft is to approve it because it reads fine.

- You have managed agencies and teams but never personally owned a number or built a system.

- You need a finished machine. This one is real and unfinished, and that is the job.