Portfolio Projects

Certificates don't get interviews; evidence does. Do 2–3 of these in order of difficulty and you'll walk into interviews with real artifacts to discuss. Do the Starter two first — they feed the later ones.

Deep product teardown of an app you love (or hate)

Starter1 week

The fastest way to start thinking like a PM and produce a public artifact recruiters can read in 5 minutes.

Product senseCompetitive analysisWriting
  1. Pick a product with obvious friction (your bank's app is usually a goldmine).
  2. Define its target user, core job-to-be-done, and business model.
  3. Map the main user journey screen by screen; note every point of friction.
  4. Compare against 2 competitors on the 3 dimensions users care about most.
  5. Propose ONE improvement: the problem, your solution, and the metric it moves.
  6. Publish as a post or PDF with screenshots.

Deliverable: A 1,000–1,500 word teardown published on LinkedIn or Medium.

10-interview user research sprint

Starter2 weeks

'Tell me about a time you talked to users' kills career-changer candidates. This project makes the question easy.

User interviewsSynthesisProblem discovery
  1. Pick a problem space you have access to (e.g., how friends manage monthly budgets).
  2. Write a discussion guide using The Mom Test rules — past behavior, no pitching.
  3. Interview 10 people; record and take verbatim notes.
  4. Synthesize into 3–5 personas or pain themes with supporting quotes.
  5. Write an opportunity brief: which pain is worth solving, for whom, and why now.

Deliverable: A research readout deck (8–10 slides) plus your interview guide.

Full PRD + clickable prototype for a new feature

Intermediate2–3 weeks

The PRD is the PM's core document. Having a real one to walk through transforms your interviews.

Spec writingPrioritizationFigma basicsMetrics
  1. Take the winning insight from a teardown or research sprint.
  2. Write the PRD: problem, evidence, goals + metrics, user stories, scope (with an explicit 'not doing' list), risks.
  3. Sketch the flow in Figma (free) — boxes and arrows beat pretty pixels.
  4. Define launch criteria and an A/B test plan for the primary metric.
  5. Get one engineer and one designer friend to review it; iterate once.

Deliverable: A 3–5 page PRD and a clickable Figma flow, linked from your portfolio.

Analytics deep-dive on a public dataset

Intermediate2 weeks

Proves you can find the story in data — the exact skill execution interviews test.

SQLFunnel analysisData storytelling
  1. Pick a public product dataset (e.g., e-commerce events on Kaggle).
  2. Define the product's north-star metric and its input metrics.
  3. Write SQL to build the funnel: acquisition → activation → retention.
  4. Find one surprising insight (a drop-off, a segment difference, a trend).
  5. Write the memo: insight, root-cause hypotheses, recommended action, expected impact.

Deliverable: A 2-page insight memo + the SQL, in a GitHub repo or notebook.

Launch a tiny real product to real users

Advanced4–6 weeks

Nothing outranks 'I shipped something and got users'. No-code tools mean anyone can do this now.

End-to-end ownershipGrowthIteration
  1. Pick a micro-problem from your research sprint — something 50 people have.
  2. Build the smallest possible version (Notion template, Tally form + sheet, no-code app, or code if you can).
  3. Get your first 20 users manually — DMs, communities, friends-of-friends.
  4. Instrument basic analytics; watch what people actually do vs say.
  5. Ship 2 iterations based on usage; write up the journey with numbers.

Deliverable: A live product link + a launch retrospective post with real metrics.

Design and simulate an A/B test end to end

Advanced1–2 weeks

Execution interviews at data-driven companies go deep on experimentation. Most candidates hand-wave; you won't.

Experiment designStatisticsDecision-making
  1. Pick a real product change (e.g., moving a paywall earlier in onboarding).
  2. Define hypothesis, primary metric, guardrail metrics, and segments.
  3. Calculate sample size and duration from realistic baseline numbers.
  4. Simulate results (spreadsheet is fine) including one ambiguous outcome.
  5. Write the decision memo: ship / iterate / kill, and defend it.

Deliverable: An experiment design doc + decision memo — a killer interview artifact.