The task

Build a small Swift app that models a real-money match from quote to settlement. A local mock server should emit match-started, score-updated, connection-lost, reconnected, and settled events; the UI must remain trustworthy through every transition.

Deliverables

1. 1A runnable Xcode project with a polished match screen, wallet summary, reconnect state, and final result.
2. 2A documented state machine covering duplicate, delayed, and out-of-order events plus app backgrounding.
3. 3Focused tests for idempotent settlement, cancellation, reconnect, and stale-event rejection.
4. 4A short architecture note explaining concurrency boundaries, persistence, and the failure you found hardest.

The task

Build a Kotlin/Compose tournament client that survives unreliable mobile connectivity. Users join a four-player bracket, play a simulated match, submit a score, and see the bracket settle from server-authored events.

Deliverables

1. 1A runnable Android project with join, match, reconnect, pending-submission, and settled-bracket states.
2. 2A fake event source that can inject latency, process death, duplicates, and events arriving in the wrong order.
3. 3Tests proving restored state cannot double-submit an entry or overwrite a newer settlement.
4. 4A concise note on lifecycle, persistence, coroutines, and what would change against a production API.

The task

Create an original 60–90 second head-to-head skill game in Unity. Both players receive the same seed and inputs are recorded so a match can be replayed and independently verified. Make the core decision loop deep enough that repeated play reveals skill.

Deliverables

1. 1A playable Unity project with a complete match loop, score, restart, and clear rules.
2. 2A deterministic replay verifier that produces the same final score from seed plus input log.
3. 3A network-condition harness demonstrating latency, jitter, packet loss, and reconnect behavior.
4. 4A design note covering sources of skill, anti-cheat boundaries, mobile performance, and three balance changes tested.

The task

Design a new synchronous, two-player mobile skill game for three-minute sessions. The winner must be determined primarily by repeatable player decisions—not randomness, spending, or device advantage—and the design should remain interesting after 100 matches.

Deliverables

1. 1A rules document with controls, scoring, match flow, onboarding, tie handling, and accessibility considerations.
2. 2A playable paper, spreadsheet, Figma, or engine prototype that demonstrates the central decision loop.
3. 3A balance model or simulation exploring at least three dominant strategies and how the rules respond to each.
4. 4A test plan with hypotheses, success metrics, sample-size reasoning, and changes made after at least three external playtests.

The task

Implement a Rust service that settles signed match results into a double-entry wallet ledger. Two score submissions may arrive concurrently, messages may be delivered more than once, and the process may crash between accepting a result and publishing settlement.

Deliverables

1. 1A runnable service with documented HTTP or message interfaces, persistent storage, and a one-command local demo.
2. 2An atomic ledger design for entry holds, payout, refund, and immutable audit history.
3. 3Tests for duplicate delivery, conflicting scores, concurrent settlement, crash recovery, and invalid signatures.
4. 4A benchmark with methodology and a brief production plan for observability, migration, and incident recovery.

The task

Build a real-time match-operations console for a fictional skill-gaming service. Operators should see live matches, inspect an event timeline, flag a suspicious result, and issue a refund with an explicit confirmation and audit trail.

Deliverables

1. 1A working React/Next.js UI and typed backend with seeded local data and a deterministic live-event simulator.
2. 2A server-authoritative event model with reconnect, pagination, optimistic-action rollback, and role-based authorization.
3. 3Tests for the refund invariant, reconnect gaps, stale mutations, and unauthorized actions.
4. 4A short system note covering data ownership, observability, and the first scaling bottleneck.

The task

Redesign the mobile flow from choosing a paid match through understanding the result and cashing out. The challenge is to make price, risk, connection state, and final settlement unmistakable without slowing down repeat players.

Deliverables

1. 1A mapped end-to-end flow including first-time, returning, insufficient-balance, disconnect, loss, win, and delayed-settlement paths.
2. 2High-fidelity responsive designs plus a coded interactive prototype of the riskiest interaction.
3. 3A compact component/token specification with accessibility states and content rules.
4. 4Research evidence from five representative usability sessions, with findings, rejected ideas, and resulting revisions.

The task

Threat-model and harden a fictional skill-game settlement API with login, wallet, score submission, WebSocket match updates, and an operator refund endpoint. Demonstrate one high-impact abuse path safely in a local environment, then design the fix.

Deliverables

1. 1A system/data-flow diagram, trust boundaries, assets, attacker profiles, and ranked abuse cases.
2. 2A minimal local proof of concept for one race, authorization, replay, or integrity flaw—never against Wagr or any live target.
3. 3A patched design or implementation with regression tests and a clear explanation of residual risk.
4. 4A prioritized 30/60/90-day security plan spanning product, cloud, secrets, detection, and incident response.

The task

Design a fraud-detection approach for a skill-gaming platform facing multi-accounting, collusion, payment abuse, and automated play. Create a synthetic event set and use it to show how your system separates suspicious behavior from strong legitimate players.

Deliverables

1. 1A reproducible generator for at least 100,000 account, device, payment, match, and withdrawal events with planted fraud rings.
2. 2A feature pipeline and detection method combining rules, graph signals, or models, with interpretable case output.
3. 3Time-based validation reporting precision/recall, expected dollars prevented, review load, and false-positive cost by cohort.
4. 4An investigation playbook and monitoring plan for drift, feedback loops, and adversarial adaptation.

Question 1

Design a model that can make Surge profitable. Surge charges an entry amount E and pays E × m(score) on a monotone score-based curve shown before play. Player skill, learning, entry-size selection, and changing game difficulty can all shift the score distribution. Estimate those effects, propose a payout curve, and prove when the economics do and do not work.

Question 2

Given a game mode where a player pays an entry e and wins e × m only after getting n wins in a row, devise a strategy that maximizes expected revenue from players. Assume every gameplay win is independent with probability 50%, and player skill is irrelevant. Derive the economics, choose n, m, and any entry or attempt structure, then analyze expected revenue, payout variance, repeat-play behavior, and the conditions under which the strategy fails.

Deliverables

1. 1A reproducible notebook or program that generates or imports player-level score histories and fits a distribution separating player skill, learning, and game difficulty.
2. 2A walk-forward backtest with player-level holdouts and sensitivity tests for adverse selection, drift, sparse new players, high-skill cohorts, and concentrated entry volume.
3. 3A proposed payout curve and entry-limit policy reporting expected margin, player RTP, confidence intervals, tail loss/CVaR, drawdown, and bankroll or ruin risk.
4. 4A two-page decision memo stating the profitability claim, strongest failure mode, kill switches, recalibration cadence, and exact production data needed next.
5. 5For Question 2, show the full derivation, compare multiple n and m choices, quantify expected revenue and payout liability, and justify the final strategy.

The task

Produce a launch-readiness plan for a fictional real-money skill game entering New York, New Jersey, and Pennsylvania. The product includes identity verification, deposits, location checks, paid contests, withdrawals, promotions, and account restrictions.

Deliverables

1. 1A sourced jurisdiction-by-jurisdiction issue matrix that distinguishes verified law, interpretation, assumption, and open counsel question.
2. 2A control map from obligation to product behavior, owner, evidence, monitoring, escalation, and release blocker.
3. 3Detailed flows for KYC/AML, sanctions, geolocation failure, self-exclusion, suspicious activity, complaints, and record retention.
4. 4A mock change-management exercise responding to one material regulatory change 48 hours before launch.