Browse and take tests to sharpen your skills
Advanced batch vs streaming design: exactly-once semantics, stream-table duality, dual pipelines, state scaling, cost/latency tradeoffs, and operational failure modes. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Choosing and operating batch vs streaming paths: windows, watermarks, delivery semantics, backpressure, CDC, and serving fresh features. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Foundations of batch and real-time data processing: latency, scheduling, use cases, and how pipelines differ. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Advanced data drift for production ML: multivariate and embedding drift, MMD and domain classifiers, concept vs covariate shift under delay, causal pitfalls, and governance of retraining. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Operational data drift detection: PSI and statistical tests, feature vs prediction drift, windowing, alert design, and retraining triggers. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Foundations of data drift in ML: covariate shift, monitoring live inputs against a reference baseline, and why production distributions change. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Advanced model registry governance: multi-region replication, model cards, supply-chain attestations, shadow/canary patterns, federated registration, and enterprise audit design. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Operational model registry practice: aliases and champions, promotion gates, signatures, access control, CI/CD handoffs, and comparing registered versions before production. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Foundations of ML model registries: versioning, stages, artifacts, lineage, and why teams centralize models instead of scattering files. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.
Sequential testing, interference, stratified bucketing, bot filtering, portfolio holdouts, cross-platform migration, and governance at scale on experimentation infrastructure.
Layers and holdouts, targeting and ramps, sticky assignment, sample ratio checks, warehouse exports, and operational hygiene on experimentation platforms. Mix of single-answer and multiple-answer items.
How A/B and feature-flag platforms assign users, log exposures, allocate traffic, and support safe rollout basics. Mix of single-answer and multiple-answer items; multi-select questions ask you to select every correct statement.