AI Architect Assessments

Scenario-driven, architect-level questions with full rationale per answer. Free, open to everyone — no login required.

LLM Architecture Fundamentals

Architect · 12 questions · 18 min

Attention, KV cache, decoding strategies, and the failure modes that only show up at production scale.

Start →

RAG & Vector Search Systems

Architect · 12 questions · 16 min

Retrieval architectures, embedding choices, ANN index trade-offs, and the failure modes hidden in real corpora.

Start →

MLOps & Production AI

Staff · 12 questions · 15 min

Deployment patterns, monitoring, drift, governance, and the on-call reality of running ML in production.

Start →

AI System Design at Scale

Architect · 12 questions · 18 min

Distributed training, multi-region inference, capacity planning, and the architectural decisions that hold up at billion-request scale.

Start →

Responsible AI, Safety & Governance

Architect · 10 questions · 14 min

Risk frameworks, evaluation, red-teaming, privacy, and the architectural patterns that make governance load-bearing.

Start →

Distributed Training & GPU Infrastructure

Architect · 12 questions · 16 min

Parallelism strategies, collective communication, checkpointing, and the GPU/network economics that decide what's actually trainable.

Start →

Agentic Systems & Tool Use

Architect · 12 questions · 14 min

Planning loops, tool schemas, memory, evaluation, and the architectural patterns that contain agent failure modes.

Start →

Recommender Systems & Ranking

Staff · 10 questions · 14 min

Candidate generation, ranking, calibration, position bias, and the eval methodology that separates real recsys teams from cosine-similarity demos.

Start →

Vector DB Internals

Architect · 10 questions · 13 min

ANN algorithms, quantization, hybrid search, and the operational reality behind production vector databases.

Start →

AI Cost Optimization

Architect · 10 questions · 14 min

GPU economics, inference cost levers, caching, quantization, distillation, and the architectural moves that reduce cost without lying about quality.

Start →

Tip

These are calibrated to a Staff/Architect bar. Choose the best answer, not just a correct-sounding one — every option here has been chosen to look plausible. Read the rationale even on questions you get right.