Product Manager

Meadow.ai · US remote · Reports to the VP of Product

Your manager is in Seattle on Pacific hours, and we'll expect some overlap. Most of our customers cluster around Houston on Central time, and there's a growing nexus of engineering and sales in the Northeast on Eastern. Live anywhere in the US that lets you be present for all three.

About Meadow

Meadow is the AI co-manager for multi-unit restaurants and retailers. We install cameras and mics in stores, fuse what they see and hear with POS and labor data, and hand operators things they can act on this week: an AI Secret Shopper that replaces clipboard audits, upsell coaching from real cashier conversations, missed-charge detection, and a BI dashboard that ties it all together. Our customers range from a handful of stores to ≈100 locations, and most are franchise systems where the person reading our data is the one running the shift.

We're an AI company and we run like one. The day-to-day work here is done with agents, not just about them. We're a small team that just raised a $7M Seed round led by Ulu Ventures, and the next 12 months are about turning a product that works into a product customers can't run their stores without.

The role

This is an execution-first product role with a customer-facing spine. You'll own the delivery cadence across our engineering teams and take over our customer syncs, so the VP of Product can spend his time on strategy, new AI capabilities, and the executive tier of those same customers. He'll join the important calls or catch the recordings; the rest are yours.

Concretely, you'll be the person who knows what's in flight on every team, why it's prioritized the way it is, and which customer is waiting on it. You'll run refinements and plannings with the engineers, keep the backlog honest, accept work before it ships, and then get on a call with the franchisee and walk them through what changed. You'll be in the data every day: reconciling a dashboard number against a POS export, catching the pipeline gap before the customer does, and turning "this looks wrong" into a ticket an engineer can pick up cold.

You'll do all of that the way we do everything here: with agents in the loop. Before a discovery call you'll have already run the customer's raw exports through an agent and come in with a proposed data mapping and a list of questions. When a franchisee asks for something, you'll bring an AI-generated prototype to the next call to find out if they actually want it before we spend engineering time on it.

We're being direct: year one is mostly running the machine and being the customer's product person. How the product surfaces get divided between you and the VP of Product isn't preset. It'll follow your skills, expertise, and interest, and the first one you own outright will show up when the cadence is under control.

What you'll do

• Run the cross-team delivery cadence: refinement and planning for FE, BE, and AI, priority communication to the teams, release notes, and the "is it on staging yet?" loop.
• Run our customer syncs after a handover: prep, data, demos of what shipped, capturing asks, following up. Roughly seven brands on a biweekly rhythm today, plus onboarding sessions for new ones.
• Pre-process a new customer's POS, loyalty, and labor exports with an agent to propose data mappings and surface the questions before the call, not after.
• Bring AI-generated prototypes into customer conversations to gauge interest before engineering time is spent.
• Be the first line on data questions, with SQL and an agent at hand. Reconcile numbers, validate before they go to a customer, and file precise tickets when something's off.
• Write the specs and tickets engineers can build from without a meeting. Acceptance criteria, edge cases, the why.
• Run acceptance on shipped work against the customer's expectation, not just the ticket.
• Coordinate launches with Ops (installs, training, comms) so a new location or feature lands as one motion, not three.
• Produce the recurring customer artifacts that ride on the product as pipelines, not by hand: biweekly insight emails, training walkthroughs, release comms. Build the ones that don't exist yet.
• Own a product surface as your skills and interest dictate. The BI dashboard is the likeliest first one.

How we work with AI

Bleeding-edge use in day-to-day work is expected here, not a bonus. Today the VP of Product's own week runs on agents that pull and reconcile customer data before a call, mine meeting notes and customer docs into a knowledge base, file tickets and open draft PRs, build the biweekly customer digest from scripted data pulls with review gates, prototype dashboard cards and decks, and run research sprints. You'll inherit some of these and build your own.

You should already have personal and professional agentic pipelines that make your life easier, and be able to show us one. What this is not: pasting a customer's email into a chatbot and forwarding the answer.

Your first 90 days

• Days 1–30: sit in every standup, refinement, and customer sync. Learn the data path from camera to dashboard. Stand up your own working setup with our agents and skills. Ship your first ten tickets.
• Days 31–60: run refinement and planning for at least one engineering team. Co-run customer syncs with the VP of Product. Pre-process one new customer's data ahead of their onboarding call.
• Days 61–90: own the cadence for all three teams and run the non-executive customer syncs. Have caught at least one data problem before a customer did, and have brought one AI-generated prototype to a customer call.

What we're looking for

• 4–7 years in product management, technical program management, or a customer-facing technical role (implementation, solutions, technical account management) where you ran the delivery loop with engineers.
• You've fronted customer calls and can hold the room when the customer says "your numbers don't match mine."
• Data-fluent and agent-fluent. Comfortable in SQL and spreadsheets, and an agent is your default first tool for the data in front of you. You direct it, you verify it, you know where it fails, and you can show us a pipeline you built for yourself.
• Writes clearly and fast. Tickets, specs, customer emails, meeting notes that people actually read.
• Organized enough to hold three teams' worth of context in your head and on the board, without becoming the bottleneck.
• Startup-stage judgment: comfortable with rough edges, knows when to escalate and when to just fix it.
• Nice to have: restaurant, hospitality, or retail operations exposure; POS, online ordering, or loyalty SaaS background; hands-on with computer vision or ML products; Jira, GraphQL, Superset, or similar tooling.

What this role is not

• Not a strategy-first PM seat in year one. If you want to set the vision for a category, this isn't the right first role at Meadow.
• Not a scrum master or a project-tracking role. You'll run cadence because you understand the product and the customer, not as a process function.
• Not customer support. Our Ops team owns support and installs; you own the product relationship.
• Not a role where AI use is optional or a novelty.

Compensation and logistics

• Base salary: $130K–$160K plus equity. Where you land in the band is context-specific, and the higher you land, the more we'll expect you to own from day one.
• Benefits. US remote. Travel to customer sites about once a quarter.
• Reports to the VP of Product. Works daily with engineers; the CTO and engineering leads are your closest partners.

How we'll interview

A short screen, an AI-native take-home built from real Meadow material (we expect you to use agents, and we grade what they can't do), a live working session where you extend it in front of us, then a half-day loop with product, engineering, operations, and the CEO. We'll tell you where you stand at every step.

How to apply

Email careers@meadow.ai with:

• The subject line "Product Manager application"
• Your resume, as a PDF
• Your LinkedIn profile, as a full URL (linkedin.com/in/yourname)
• In 150 words or fewer: an agentic workflow you run for yourself, what it does, and one time it got something wrong
• The word platypus somewhere in the body, so we know you read all the way to the end

No cover letter needed.

Note: Only applications sent to that email address will be considered; sending them to other email addresses will not reach the inbox.