Product Manager

About the role

We build software for commercial real estate lenders, loan origination, servicing, document extraction and financial analysis.

The product is AI-heavy, and so is the way we build it.
 

This role owns the delivery cycle: sprint planning, requirements on the wiki, tickets, testing, design review. What makes it different is the method. You will not write forty tickets by hand or click through regression tests manually. You will build and run agents that do the first pass, and your value is in the specification you give them and the judgment you apply to what comes back.
 

We are looking for someone who has already worked this way, or who will obviously thrive doing so.
 

Schedule: read this first
This role bridges a development team in Europe and a business team in the US. Two fixed daily commitments:
10:00–12:00 Kyiv time — morning call with the dev team, Monday–Friday
 

19:00–21:00 Kyiv time — call with the US team, Monday–Friday
 

Both are fixed, five days a week. The hours between are yours to structure, and the split is real, plan for it rather than discovering it in month two. Apply only if this works with your life long term.

What you'll own
- Sprint management
- Create and run sprints; manage planning, scope and capacity
- Keep the backlog groomed and prioritized with engineering and leadership
- Requirements and documentation
- Write feature descriptions and specifications on the wiki
- Keep wiki, designs and tickets consistent with each other
- Tickets
- Create Stories and Bugs in the team's required format, with clear acceptance criteria
- Build reusable agent workflows that produce tickets to standard instead of drafting each from scratch
- Testing
- Write and execute test plans; run regression passes before release
- QA AI extraction output against source documents and file defects with the source evidence
- Design
- Review implementation against Figma designs and flag mismatches before they reach release
- Catch gaps between design, spec and built behavior
- Product development
- Participate in scoping and design discussions; represent the user in build decisions


What we're looking for
-3+ years in product management, delivery management, or a PM/QA 
-AI is already how you work, not a tool you occasionally open, you can walk us through your daily setup and what it replaced
-Hands-on with agent workflows: Claude Skills, MCP / tool use, Claude Code, Cowork or comparable frameworks
-You can write a specification precise enough that an agent produces usable output from it — the same skill that makes a good ticket
-Judgment about where AI output needs verification, and the discipline to actually verify it
-Comfortable with messy documents and structured data, spreadsheets, PDFs, extraction schemas
-Strong spoken and written English and Russian, you run calls in both, daily
-Able to commit to the fixed call schedule above on an ongoing basis
Familiarity with Jira, Confluence and Figma
 

Nice to have
-Commercial real estate, lending, fintech or insurance domain experience
-QA background, manual or automated
-SQL, or enough Python to inspect data and prototype
-Experience with document-heavy B2B workflows
 

How we work
-Tooling is paid for and expected to be used daily
-Small teams, short release cycles, direct access to customers
-Two fixed daily calls (see Schedule); everything else is async and flexible
-You are the connection point between the dev team and the US business team
 

What we offer
-[Employment type — contract / EOR / full-time]
-[Benefits, equipment budget, learning budget]