Product / Program Manager · Software for teams & AI product experiences

View my portfolio and resume · Browse the guided demos

I work on software products for teams: understanding a problem, choosing what to improve, defining how it should work, and checking whether the result is useful. This portfolio makes those decisions visible through fifteen interactive demos, case studies and product requirements documents (PRDs).

My contribution: product problem framing, prioritization and tradeoffs, requirements and workflows, fictional sample-data design, success criteria, and evaluation. AI tools assisted with drafting, implementation and verification. The product decisions, requirements and evidence are the basis for reviewing my work. Each demo has deliberately narrow scope so a reviewer can inspect the choice, its cost and what would change the direction.

My background spans recommendation and onboarding work at Oracle, enterprise search and asset review at 5th Kind, mobile growth and packaging at Mabrosi, and program delivery at Raytheon. These demos are independent fictional exercises in those product decisions. Each case study explains the connection and a question that could challenge the chosen approach.

• AI product judgment: Relevance Studio case study — decide when a recommendation is useful enough to show and when to admit there is no fit. Try it.
• Discovery and search: Findability Lab case study — distinguish missing metadata from a search-policy problem before investing in a more complex solution. Try it.
• Program leadership: Program Atlas case study — explain the impact of a delay, compare recovery costs and identify who must accept the remaining risk. Try it.

For a quick onboarding example, First Mile demonstrates value before complete setup. All demos use fictional records and run in your browser. The walkthroughs give exact steps and expected results.

You can review the portfolio without reading code. Start with a project’s case study for the decision and tradeoff. Follow its sample walkthrough to reproduce a concrete scenario, then use the PRD and validation plan to inspect the rules and evidence limits.

For Evaluation/Evidence, start with Experiment Verdict; for Value Realization, start with Outcome Ledger. For Lifecycle/Growth, start with Return Path; for Enterprise/Account Strategy, start with Renewal Compass. For Search/Metadata, start with Findability Lab; for Packaging/Monetization, start with Package Forge. Relevance Studio covers AI/Relevance, and Review Room covers Enterprise/Collaboration. First Mile covers Product/Growth, and Program Atlas covers Program/Value. Feedback Nexus shows evidence-based prioritization, while Task Assistant shows confirmation and recovery boundaries. The walkthroughs use exact sample inputs and observable results; they are product tutorials, not claims of customer validation.

The portfolio connects product discovery, onboarding, AI and search policy, enterprise content review, packaging, lifecycle engagement, account strategy, experiment evaluation, benefit evidence, and program delivery. Each case makes a distinct decision inspectable through requirements, a working sample, and recovery.

Northstar gives the examples a consistent fictional setting, but the apps are independent: a shortlisted feedback item does not create an Asana task, change a staffing plan, or approve a launch. Each demo makes one decision and its limits inspectable. Together they show how I define useful scope, preserve human judgment, and decide what evidence is missing before investing further.

The portfolio's decision questions remain relevant as organizations adopt agents. Microsoft's 2026 Work Trend Index emphasizes human agency, judgment and organizational readiness. These themes support examining who makes a decision and whether the surrounding workflow is ready for it.

For 2027 planning, my proposed investment lens is to ask who can authorize an action, which context deserves trust, what useful outcome would justify the operating cost, and when a simpler workflow should win. These are planning hypotheses to test. The demos expose the product questions with fictional data; they do not demonstrate deployed agents or future market outcomes.

Choose the question you want to evaluate. Each document uses the fictional Northstar scenario and makes its assumptions explicit.

These are product artifacts for review, not records of completed customer research or shipped commercial outcomes. The requirements and demos show what can be inspected now; the discovery, market, and evaluation plans explain what I would need to learn before investing further.

Each case study distinguishes a product decision, a demonstrated sample behavior, and an unanswered research question. The walkthroughs make the sample behavior reproducible; the measurement plans define what would count as evidence before further investment.

Northstar is a fictional B2B SaaS company used across these independent examples. The demos require no login, paid API, real customer data, or live external integrations. Task Assistant uses predefined rules entirely in the browser; it has no AI model, server backend, or Asana connection. Its existing Asana-Agent URL is a legacy address.

Proposed success metrics are hypotheses, not achieved business results. Software checks establish whether specified behaviors work; they do not replace customer research, demonstrate adoption, or prove commercial impact. Validation plans distinguish those open questions from the behavior demonstrated today.

Source code remains public for transparency. The PRDs, case studies, and interactive workflows are the starting points for evaluating my product management work.