Work

Product decisions made under real pressure.

My work sits where customer needs, technical systems, operating realities, and organizational decisions meet. These stories show how I frame the problem, make the tradeoffs visible, and move teams toward a useful outcome.

Selected project stories

Three different environments. The same responsibility.

Create clarity, protect what matters, and design a path that people and systems can actually support. Commercially sensitive details are intentionally generalized.

Al Jazeera Media Network

2024 - Present · Trustworthy AI product

Ask Al Jazeera: conversational access to verified journalism

A reader-facing experience that helps people ask questions and explore the news through answers grounded in verified Al Jazeera reporting.

Situation

Readers often arrive with a specific question, while the answer may be distributed across a large and continuously evolving body of reporting. Traditional search can surface documents without resolving the reader's underlying need.

What made it difficult

A useful answer also has to be trustworthy. Retrieval quality, source attribution, freshness, editorial standards, model behavior, and the path back to the original reporting all matter as much as fluency.

My role

I led the product direction within Al Jazeera's broader AI strategy and AI Lab, connecting editorial needs with design, engineering, data, and AI specialists. I framed the experience, evaluation approach, governance boundaries, and path from prototype to product.

Key product decisions

Ground answers in verified Al Jazeera sources rather than open-ended model knowledge.
Make source transparency part of the experience so readers can inspect the reporting behind an answer.
Use structured evaluation and prompt governance to test usefulness, accuracy, and failure modes before scaling.

Maqsam

2021 - 2024 · SaaS analytics

Turning call-center data into decisions

A reporting engine and KPI calculator that made a cloud communications platform more useful to the administrators responsible for call-center performance.

Situation

Call-center administrators needed to understand queues, agents, activity, and service performance without translating raw interaction data into operational meaning by hand.

What made it difficult

The same metric can mean different things across teams, time windows, and operating models. The product needed calculation logic that was consistent enough to trust and flexible enough to reflect how different customers ran their centers.

My role

I led discovery with customers and internal teams, shaped the reporting model and KPI language, prioritized the product sequence, and aligned product, design, engineering, data, sales, and customer-facing teams around what administrators actually needed to decide.

Key product decisions

Start with the administrators' recurring questions, not with the fields available in the database.
Create explicit KPI definitions and reusable calculation logic so the same measure stays consistent across views.
Design the reporting layer for comparison and investigation, not just static summaries.

Expedia Group

2017 - 2021 · Operational automation

Automating airline refunds when the system was under pressure

An automated refund process designed during COVID-19, when disruption drove exceptional volumes and manual operations could no longer absorb the demand.

Situation

Travel disruption produced a surge of refund requests at the same moment customers needed speed and reassurance. Existing manual steps created workload, delay, and operational risk across the airline e-commerce experience.

What made it difficult

Refunds crossed payment flows, airline rules, booking states, operational teams, and exception paths. The solution had to move quickly without treating every case as identical or hiding failures inside an automated process.

My role

I led the product work across the end-to-end refund journey, mapping the workflow and dependencies, aligning operations, finance, customer support, engineering, and platform stakeholders, and defining where automation was credible and where human judgment still mattered.

Key product decisions

Standardize the decision logic before automating the workflow around it.
Separate straight-through refund cases from exceptions that required investigation.
Build operational visibility into the process so teams could understand status, failures, and handoffs under pressure.

Broader portfolio

Product leadership across the surrounding system.

The visible feature is only one layer. The product function, platform, data, payments, and operating model often determine whether it can create value.

AI Lab and newsroom workflows

Founded and led Al Jazeera's AI Lab, introducing governed GenAI workflows for writing, translation, media search, and content packaging alongside evaluation, prompt governance, and model-lifecycle practices.

Building the product function

Established the product management discipline at Maqsam from inception, creating the practices, decision rhythms, and team foundations needed to grow the function.

Payments and platform reliability

Used analytics to reduce payment failures and led Maqsam's migration to Stripe, improving payment reliability and the prepaid-credit experience for customers.

Commerce, APIs, and cloud evolution

Helped shape Expedia's B2C and B2B airline commerce experiences, API-platform direction, and transition from traditional infrastructure toward AWS-based services.

Conversation

Working through a product problem with similar complexity?

Start a conversation