Haystack
← Back to Jobs
Operations & Project Management
TA

AI Product Manager

TalTeamUnited States🇺🇸United StatesPosted Sep 18, 2026

Quick Overview

Seniority
Mid Senior
Work mode
On Site
Location
United States
Posted
Yesterday

Job Description

Position : Product Manager
Location : Burlington MA - Remote
Direct Client Requirement
Job Description:
We are building our first generation of customer-facing AI experiences, and we need a product manager to lead them from early concept to something customers rely on, then keep going.
This is a zero-to-one role inside a large, regulated enterprise, on a team that is being built from the ground up. There is no established playbook, no mature AI platform team, and no roadmap waiting for you. You'll build it, and you'll have a real hand in shaping how the team works, what it prioritizes, and who joins it next.
Regulated, here, means concrete product constraints: hazardous materials, export
controls, and scientific accuracy where a wrong answer isn't just unhelpful. Designing AI that customers can trust in that environment is part of the craft, not a box to check.

What you'll own
Day one. An on-site AI shopping assistance. You'll define what "good " means, instrument it, and iterate quickly with engineering, data science, and UX.
From there. The frontier. Build what's next with teams across the business.
Create and scale AI capabilities, from agentic purchasing and personalization to generative AI across the customer journey that help teams deliver better outcomes.
You'll identify opportunities, run experiments, learn quickly, and shape a roadmap grounded in customer and business impact.
From there. The frontier. Build what's next with teams across the business.
Create and scale AI capabilities, from agentic purchasing and personalization to generative AI across the customer journey that help teams deliver better outcomes.
You'll identify opportunities, run experiments, learn quickly, and shape a roadmap grounded in customer and business impact.
Infrastructure and Financial Ops. The product leadership side of model and
vendor selection, data access, and LLM spend management, so the work stays fast, reliable, and aVordable as it grows.

Quality
Ensuring we ship work that is useful, valuable, and delightful to our customers.
That includes how retrieval, prompting, and guardrails shape what customers see, and the evaluation

How you'll work
The fundamentals are yours to run, not delegate:
Take products from zero-to-one: turn ambiguous problems into a sharp vision,
testable hypotheses, and a first version customer can use.
Own the backlog end to end: epics, user stories, acceptance criteria, and ruthless
prioritization that keeps the team unblocked.
Run discovery and write the specs: talk to customers, dig into usage data, and turn
what you learn into experiment plans engineers and designers can build from.
Define the metrics, instrument them, and report results plainly, including the
experiments that failed.

What success looks like at 12 months
Customer-facing AI shopping experiences are live, adopted by a meaningful share of
visitors, and measurably improving customer outcomes such as task completion,
findability, and repeat visits.
A working evaluation and measurement loop exists, so the team can tell whether a
change made the product better before customers tell us.
At least one net-new AI experience beyond conversational assistance has shipped
to real customers, and we've learned something concrete from the ones that didn't.
The team has a clear, repeatable path to the data it needs, and a well-argued case
for the data it doesn't have yet.
Engineering, data science, UX, search, and legal partners see you as the person who made AI work here.

Who you'll work with
Engineering, data science, UX, search, legal and compliance, IT, and commercial teams.
You'll be in the room with engineers and designers daily, you will work hand-in-hand with them every single day. This is a hands-on partnership role, not a coordination role.

What you bring
5+ years in product management, with at least one LLM powered product you shipped to real users. Scale matters less than ownership, a side project people rely
on counts, and so does an internal tool or a small feature at a large company. What
we want to hear is what you decided, what broke, what you measured, and what you
changed.
Strong core PM craft: backlog ownership, epic and story writing, prioritization,
specs, discovery, and metrics. You've run a delivery team's day-to-day and can
show the artifacts.
Working fluency in the AI landscape: foundation models, major vendors, retrieval,
evals, latency, cost. You can hold a technical conversation with engineers and
challenge assumptions.
Hands-on evaluation skills. You've written evals, read test results, and made ship
decisions from them.
You use AI tools daily to prototype, analyze, and write, and you can show us
something you built with them.
Resourcefulness with imperfect data. Our data lives in many systems, some of them
old, and access isn't always fast. You know how to ship valuable experiences with
what's reachable today while building the case for what you need next.
A builder's disposition. You prototype, you test with users, you'd rather ship
something small this week than plan something large this quarter.
Comfort operating inside a large enterprise: navigating legal, compliance, IT, and
stakeholder alignment without letting it slow the work to a crawl.
Strong judgment about customer value. You can tell the diVerence between an AI
feature that's impressive and one that makes a scientist's day easier.

Nice to have
eCommerce, B2B eCommerce, marketplace, or catalog-heavy product experience
Life sciences or scientific customer exposure
Experience with personalization, search, or recommendation systems
Experience working with ERP, PIM, or other enterprise systems of record as data sources
Founder or early-stage startup background

Similar jobs