Product Manager - AI products
Pyramid AI · Hong Kong · Reports to Head of Product
About Pyramid AI
Pyramid AI is building the intelligence layer for construction workflows.
Construction professionals spend enormous amounts of time interpreting drawings, specifications, requirements, and domain knowledge to check, extract, compare, and produce work. We believe AI can fundamentally change how this work gets done.
The Role
We’re looking for a high-agency Product Manager to own products from discovery through customer value.
You’ll identify important customer problems, develop and validate product theses, work closely with engineering to build solutions, put products in customers’ hands, measure whether they create value, and iterate until customers adopt and pay for them.
Your immediate focus will be our Design Compliance & Studio module.
Job Description
You’ll own the product lifecycle end to end—from identifying a valuable customer problem to delivering a product customers adopt and pay for.
Discover → Form a thesis → Validate → Prototype → Build → Ship → Measure → Iterate
This is a hands-on ownership role. You won’t simply gather requirements, write specifications, or hand work to engineering. You’ll remain accountable for the quality of the thinking, the pace of execution, and the customer outcome.
1. Discover valuable problems
- Lead customer interviews, workflow observations, and market research
- Understand how customers work today, where time and money are lost, and why the problem matters
- Separate what customers ask for from the underlying need
- Prioritize problems based on customer value, strategic fit, and commercial potential
2. Form and validate a product thesis
- Turn incomplete and sometimes conflicting evidence into a clear point of view
- Define who has the problem, how they solve it today, what should change, why Pyramid can solve it better, and what must be validated
- Identify the riskiest assumptions and design the fastest way to test them
- Use prototypes, demo workflows, AI-generated interfaces, and lightweight proofs of concept to learn before committing to a full build
3. Drive the smallest valuable release
- Translate the product thesis into clear outcomes, scope, and success criteria
- Work side-by-side with engineering and domain experts so the team understands both the customer context and the why
- Make aggressive trade-offs, remove ambiguity and blockers, and keep the team focused on the smallest release that can create or validate customer value
- Get working software in front of customers early rather than waiting for a perfect launch
4. Own the outcome after shipping
- Observe real usage and gather structured customer feedback
- Measure whether the product solves the problem, earns repeat use, and creates willingness to pay
- Turn evidence into clear decisions: iterate, expand, reposition, or stop
- Build toward a repeatable product—not a collection of bespoke customer requests
Your goal is to shorten the loop from customer problem to measurable customer value without sacrificing the quality of the product thinking.
How We’ll Measure Success
Your primary metric is Time to Customer Value: how quickly we move from identifying a valuable customer problem to delivering something that creates measurable value.
Depending on the product stage, success may include:
- Time to first usable prototype
- Time to first customer use
- Adoption and repeat usage
- Trial-to-paid conversion
- Paid customers and revenue
About You
You’re likely to be a strong fit if you:
- Have strong product judgment and can form a point of view with incomplete information
- Love talking to users and uncovering what sits beneath their initial requests
- Can turn messy customer and market information into a clear product thesis
- Think commercially and care about customer value and willingness to pay
- Communicate customer context clearly so engineers understand why they’re building something
- Know how to say no and make aggressive scope trade-offs
- Prefer prototypes and experiments over lengthy requirements documents
- Move quickly and are comfortable with daily and weekly iteration
- Take ownership beyond launch and care whether customers adopt what you build
- Are comfortable challenging founders, customers, and engineers
- Keep up with modern AI capabilities and understand how they change what’s possible in a product
You don’t need to build the underlying AI yourself, but you should be conversant in modern AI products—including LLMs, agents, RAG, multimodal models, evaluations, and emerging AI product patterns—and know enough to work productively with engineers.
Bonus Points
- You’ve shipped AI products before
- You’re comfortable using AI coding tools or a CLI to prototype ideas
- You can turn an idea into a rough working demo without waiting for engineering
- You’ve built complex B2B or enterprise products
- You’ve worked with document-heavy professional workflows
- You have experience in construction, architecture, or engineering
This Role May Not Be for You If
- You see product management primarily as gathering requirements and translating them into tickets
- You only measure progress by backlog completion or the number of features shipped
- You need everything defined before you start
- You default to building whatever customers request
- You hand requirements to engineering and move on
- You believe the job is finished when the product launches