Available for new projects
Let's talk
delivery.
Open to Project Manager roles and consulting engagements. Based in Islamabad, working across time zones.
Typically responds within 24 hours
Product Owner / Project Manager with 4+ years of experience defining product requirements, authoring BRDs, and wireframing scalable solutions across Web, Mobile, SaaS, and AI platforms. Proven track record of taking complex, ambiguous client visions and translating them into clear acceptance-criteria frameworks, achieving a 95% feature sign-off rate. Expert at backlog grooming, balancing technical constraints with user experience, and managing full-lifecycle feature delivery.
Five real projects, client details anonymized. Each one hit a real problem — technical, financial, or human — before it shipped.
An Ontario law firm needed a platform that captures a client's situation once, then automatically determines which of ~25 family court forms apply — using a decision-tree engine, not a checklist — and auto-populates them from OCR'd bank statements.
The wall: the Data Science team hit real accuracy problems getting OCR to correctly read scanned bank and credit card statements — a serious blocker, since the whole pitch depended on trustworthy financial data extraction.
The fix: worked through image preprocessing with the DS team to clean up scans before OCR, resolving it ahead of launch. Ran daily standups and brought in beta testers early to validate the AI matching logic.
Peaceful Prep helps separated parents build structured parenting plans, run AI-assisted conflict-resolution rounds, and connect with mediators — a nearly year-long build with 10 people across full SDLC ownership.
The wall: in the early stages, the accuracy of the AI's outputs — conflict summaries, agreement comparisons — wasn't good enough. In a product dealing with custody and mediation, a wrong AI suggestion isn't a minor bug.
The fix: designed a formal acceptance-criteria framework shared by both the team and the client, turning a subjective "does this feel right" problem into a testable, agreed-upon standard.
Nimble is a 20-month, web + mobile platform for construction teams: auto-scan floor plans from a phone camera, hand-draw freeform walls, attach custom forms to floors and rooms. 15-person team across Data Science, mobile, web, QA, and design.
The wall: wall-snapping ("mating") logic for freeform floor-plan drawing on mobile turned out to be a genuinely hard geometry problem. Every third-party library the team tried hit a blocker.
The crisis: mid-project, frustrated by the pace this caused, the client requested a refund.
The fix: met the client directly and walked him through the actual technical reality — not just "it's hard," but the specific blockers and why off-the-shelf libraries couldn't solve it. He agreed to continue. The team then built a custom in-house geometry solution from scratch.
Viral Velocity scores photos 0–100 across composition, emotional resonance, storytelling, and color psychology to help users decide what to post — plus AI enhancement that generates and re-scores up to 5 improved versions. Full SDLC ownership, 6-person team, 5 releases.
No major fires on this one — which is its own kind of proof. Clean scope, steady execution, tight team across backend, mobile, AI enhancement, and admin tooling.
Audit Room manages financial statement components, staff charging rates, and working papers for audit firms — with a role hierarchy from Super Admin down to individual staff members, tracking manpower across hundreds of audits. Full lifecycle ownership: requirements, wireframes, BRD, team of 10, still ~50% through the build.
The wall: the client grew impatient with the pace of a large, ongoing build.
The fix: walked him through the actual project timeline and why it was structured the way it was — restoring confidence without cutting scope or rushing the architecture.
Every project starts with a BRD and wireframes before a single sprint is planned — no exceptions, regardless of client size. Sprint plans and risk logs are shared with the client, not just the team, so nobody is surprised by a delay. When a technical blocker or a client's confidence hits a wall, I don't wait it out: I get on a call, show the actual problem (not a summary of it), and put a revised number on the table. That's what got a construction client to keep a project instead of asking for a refund, and it's what a formal acceptance-criteria process did for an AI product that was hard to spec by feel alone.
Day to day: daily standups, weekly client demos, and a single source of truth for scope — usually Jira or ClickUp, whatever the client's team already runs on.
Manual and regression testing for enterprise applications; helped shift part of the suite to automation.
Coordinated 10 parallel software projects, owning BRDs, planning, and release schedules.
Own end-to-end delivery of 30–35 projects across web, mobile, SaaS, AI and Shopify.
Formalizing a practice already built around scope control, risk registers, and structured delivery.
Open to Project Manager roles and consulting engagements. Based in Islamabad, working across time zones.