SIDE PROJECT · CURRENT

CertArena

An AI certification exam simulator and benchmark with Pearson VUE timing, Kokoro speech synthesis, and cited documentation across nine certification tracks.

9 tracks (228 items) VUE test engine scaled ≥ 720? review + citations 1,132 KOKORO AUDIO CLIPS · ZERO DEPENDENCIES
9 TRACKS · 228 ITEMS

CCDV-F, AWS AI Practitioner, Azure AI, Google GenAI, Anthropic Claude, and specialist benchmarks.

PEARSON VUE HARNESS

120-minute countdown, 720 scaled passing threshold, proctor calculator, review drawer, question strikethrough.

KOKORO SPEECH ENGINE

1,132 synthesized audio clips. Interactive read-aloud pill with live phrase highlighting.

ZERO NPM DEPENDENCIES

Native Node 24 ESM runtime. Built solely on node:http, node:sqlite, and node:crypto.

THE REASONING

Architecture map of CertArena: the real components and how they connect
THE ARCHITECTURE MAP
Zero-dependency Node 24 ESM runtime

Modern web applications carry hundreds of megabytes of third-party package dependencies for basic tasks like routing, database queries, and crypto. That complexity introduces constant supply chain vulnerability churn, security patches, and deployment friction.

CertArena ships with zero external npm dependencies. The entire system runs directly on Node 24 native capabilities: node:http handles request routing and static asset delivery, node:sqlite manages state and telemetry persistence, and node:crypto handles secure token and session generation. Cold boot takes under 80 milliseconds and runtime memory remains below 40 megabytes.

Pearson VUE exam simulation and scaled scoring

Studying for professional engineering certifications requires realistic exam conditioning, not just multiple-choice flashcards. High-stakes exams test endurance, pacing, question triage, and elimination strategies under time pressure.

CertArena replicates the Pearson VUE exam environment: a 120-minute persistent countdown timer, right-click question strikethrough, question flagging, a proctor-style floating calculator, and a full review drawer. Scoring uses the official 100 to 1,000 scaled score model with a 720 passing threshold, calculating domain-weighted scores across all syllabus competencies.

Local Kokoro neural speech synthesis and word alignment

Reading complex multi-paragraph technical scenarios creates cognitive fatigue. An audio track lets candidates absorb question stems and detailed answer explanations hands-free during long study blocks.

Instead of relying on costly third-party cloud speech APIs, CertArena leverages the Kokoro open-weights neural text-to-speech model. The pipeline pre-synthesized 1,132 synchronized audio clips spanning all question stems and comprehensive explanations. An inline listen button allows instant playback with visual sentence-level highlighting directly in the question card.

Deterministic answer shuffling and session verification

Memorizing answer letter positions degrades study value and invalidates benchmark integrity. True mastery requires evaluating choices independently of their presentation order.

CertArena implements deterministic answer shuffling. When an exam session begins, the server seeds a pseudo-random permutation for each question. The client renders options according to this session seed, and answer grading validates selections against the server-side seed map. Test-takers cannot game answers by memorizing positional keys.

Vendor documentation citations and domain diagnostics

A score number without actionable feedback does not help an engineer pass. When an answer is wrong, the student needs to know exactly which architectural concept failed and where the vendor documentation defines the correct standard.

Every CertArena question links to official vendor documentation from AWS, Azure, Google Cloud, or Anthropic. The post-exam diagnostic breaks down score percentages across individual blueprint domains, highlighting precise knowledge gaps before the candidate sits for the real exam.