Koast.ai: Head of Product @ Koast.ai

UTTR: Project Manager

September 9, 2026

Join the VA Police for a career in courage and caring

September 10, 2026

Published by Coupea on September 9, 2026

The short version

Koast is the AI operating layer for teams that run Meta ads at volume. Agencies, in-house media buying teams, and enterprise buyers managing 10+ ad accounts and launching 50+ ads a week use Koast to launch faster, automate optimizations off real attribution data, and see every account from one hub. The Koast Agent is the next layer: an AI that reads the account, spots what matters, and acts inside guardrails.

We are a small team shipping fast. The product is live, customers are paying, a Whop partnership just onboarded a wave of new users, and the Agent build is underway. What we do not have is one person who owns the product end to end: what gets built, why, in what order, and whether it actually landed.

That is this role. It starts at 20 hours a week on contract. If you are the right person, it becomes the full-time product seat at Koast.

What you will own

Discovery that finds the real problem. You will sit in on enterprise sales calls and onboarding sessions, run discovery with agencies and media buying teams, and figure out what they are actually trying to get done versus what they say they want. Every feature request gets mapped to a job: launching smoother, getting their team in, getting tasks done, keeping users active and growing spend inside Koast. If it does not serve a job, it does not get built.

Specs that engineering can execute without a meeting. We run a lean engineering team plus an AI-assisted dev pipeline. That only works when the ticket is unambiguous. You will write PRDs and Shortcut tickets that define the behavior, the edge cases, the data dependencies (Meta API limits, Hyros and Cometly attribution feeds, Whop account provisioning), and the acceptance criteria. If an engineer or an agent has to guess, the spec was not done.

The queue, and what is missing from it. You own the Up Next column. You will keep engineering fed so they are never blocked, and you will keep a standing view of what is discussed in Slack, support, and QA that does not yet have a ticket. Gaps are your problem before they become customer problems.

Evidence over opinion. PostHog, Intercom, session replays, rage clicks, support threads. You will run the user-behavior audit, know where people drop off in launching, and bring numbers to every prioritization argument, including the ones with the CEO.

Delivery from spec to shipped. Governance, milestones, release notes, and the customer conversation after it ships. When something breaks or slips, you are the one who says so early, not the one who explains it after.

Technical judgement in a non-deterministic product. The Koast Agent makes decisions off data that is noisy, seasonal, and shaped by Meta’s learning phase. You will make product calls that respect that: what the Agent is allowed to do, what it must never do, how we measure whether an automated action was correct. Guardrails are infrastructure here, not UI. You need to understand why.

What you will do week to week

Run discovery sessions with prospects and customers and turn them into scoped, execution-ready PRDs.
Surface technical risk early: undocumented API behavior, missing data, an assumption that will not survive a real ad account.
Write and groom Shortcut tickets with enough detail for an engineer or Claude Code to ship without a follow-up question.
Partner with engineering on feasibility, sequencing, and cutting scope to what ships this week.
Own success metrics per feature and report on them after launch, not before.
Be the customer’s point of contact through onboarding and rollout for enterprise accounts.
Write the updates: one version for the CEO, one for the customer, one for the engineer. Each one short.
Say “we should not build this” when the data says so.

What sets you apart

You have run Meta ads or sat close enough to a media buyer to know what a 2am budget change feels like. You know what Advantage+ does to a structure, why learning phase resets matter, and why “just pause the losers” is a bad automation.
You have shipped B2B software end to end in a small team. Not managed it from a distance. Shipped it.
You have worked on or with AI or ML products where the output is not deterministic and you had to define what “correct” means.
Senior engineers take you seriously because you read the API docs before the meeting.
You write tightly. Your PRDs are shorter than everyone else’s and clearer.
You are comfortable being the only product person in the room and the one who gets blamed when the roadmap is wrong.

Requirements

4+ years in product management on B2B SaaS, with full-lifecycle ownership from discovery through iteration.
Hands-on familiarity with Meta Ads Manager and the Marketing API. Agency or performance marketing background strongly preferred.
Strong technical fluency: you can read an API response, reason about data models, and understand why a rate limit changes the feature design.
Experience working with attribution or analytics data (Hyros, Cometly, Triple Whale, Northbeam, or similar) is a plus.
Experience writing specs consumed by AI coding agents is a plus. Willingness to learn it is required.
Comfortable in Shortcut, PostHog, Intercom, Slack, and Notion, or able to get there in a week.
Available 20 hours a week during US business hours, with the intent to go full-time.

How You Work

Small team, high trust, no theater. Decisions are made with numbers, written down, and revisited when the numbers change. These are the values we run on. If they sound like you, you will fit right in.

Mastery. “Good enough” is where you start, not where you stop. You keep sharpening the craft: tighter specs, sharper questions, cleaner calls.
Push the Boundaries. You take on the ambitious version of the problem. Autonomous media buying does not exist yet at the level we are building it. You want to be the one who defines it.
Embrace the Challenge. When a launch flow breaks for a customer at 9pm, you are in the thread, not waiting for the morning standup.
Responsibility. You own your commitments. If a date slips, everyone hears it from you first, with a new date attached.
Dedication. You dive into the hard, unglamorous problems (data gaps, undocumented API behavior, edge cases nobody wants to spec) and iterate until they are solved.
Collaboration. You write clearly, share what you learn, and make engineers, sales, and customers smarter for having worked with you.

Why Koast

Early-stage impact. You will be the first product seat. The roadmap, the spec standard, and the way this company decides what to build are yours to define.
AI and ads at scale. You will work on the frontier of AI-driven media buying, shaping how an agent makes real budget decisions for some of the highest-volume ad teams on Meta.
Growth and ownership. Start at 20 hours a week on contract. Grow into the full-time Head of Product seat, with the potential for equity.
Remote and flexible. Work from anywhere, on your schedule, during US hours. Just ship product that customers feel.

How to apply

Record a Loom, five minutes or less, covering these four things in order:

Show us an ad account. Screen-share any Meta ad account you have or have had access to (blur what you need). Tell us the last change you made in it and why.
One thing you would kill. A feature you shipped that you would remove today, and what you learned from it.
One spec. Screen-share a PRD or ticket you wrote that an engineer shipped without asking a follow-up question. Walk us through how you wrote it.
Where Koast fits. Sixty seconds on where a tool like Koast changes a media buyer’s week, and where it does not.

Email the Loom link and your LinkedIn or resume to jobs@koast.ai with the subject line “Scale with Koast“.

No cover letters. We watch the Loom first and read the resume second. Applications without the video are not reviewed.

To apply: https://weworkremotely.com/remote-jobs/koast-ai-head-of-product-koast-ai