Six years inside a B2B field sales SaaS company: first on the phone with the users, then guarding the release gate. I know how integrations fail in production because finding them before release was my job.
Case studies on real production work, a self-directed prototype I designed and shipped, and the core documents a PM actually writes: a spec and the experiment that proves it.
A solar proposal is a priced quote handed to a homeowner at the kitchen table, assembled from hardware data, live pricing, and a third-party lender API. How I automated validation for it, and the silent pricing defects that surfaced.
Field sales apps lose 40–60% of new users before their first week ends. A product brief proposing a "Quick-Log First" onboarding redesign that delivers value in three minutes, before any configuration.
A knitting problem I have myself: at the store, I couldn't remember which needle pieces I owned, and bought duplicates. The fix turned "what do I own?" into "what can I build right now?" A case study, plus a working, installable prototype.
The core document a PM produces. A full spec for a feature that tells field sales reps who to visit next and why: problem, goals and non-goals, a transparent scoring model, prioritized requirements, edge cases, and rollout.
The other half of owning a feature. A metrics and A/B plan for Next Best Visit: choosing a North Star over a vanity count, setting guardrails, and committing to a decision rule in writing before the data arrives.
I've spent six years at SalesRabbit standing at the two points where a product meets reality: the support queue and the release gate.
On support, I handled 6,900+ tickets. You learn something specific from that volume: not just what users complain about, but the shape of the complaints. Which frustrations are noise and which are the product telling you something. I turned those patterns into structured feedback that reached the roadmap.
In QA, I moved upstream. I built a bug escalation program that validated every incoming ticket before it reached engineering, cutting invalid escalations by 65%. I wrote automated regression scripts in JavaScript inside a CI/CD pipeline. And I spent a lot of time in the place I find most interesting: the integration layer, where our product depends on someone else's API behaving the way their docs promised.
That's the work I want to keep doing as a PM. Integrations are where product strategy stops being abstract. You're negotiating with a partner's rate limits, someone else's data model, and a user who doesn't care whose fault it is. I like that problem.
Support tickets, session data, and reproduction steps beat opinions. I start where the friction actually shows up.
Most failures live between systems, teams, or steps, not inside them. I look at handoffs first.
Define the metric and the guardrail before launch, or you've shipped a guess with a release date.
Open to product management and technical program management roles. Happy to walk through any of this work in more depth.