Deployment
130 bites tagged Deployment — interview questions with model answers, and 60-second explainers.
Why use an ASGI server like Uvicorn instead of the dev server?
Tests: dev vs prod environment distinction. Answer: Uvicorn is a prod server program; contrast dev server's constant restart/break/fix cycle with prod needs for performance, stability, and uninterrupted access.
FastAPI Container Build and Deploy Pipeline
Treat the Docker image as the immutable artifact: one build runs everywhere. Deploy FastAPI workers behind a load balancer, one process per container. The footgun is baking secrets into the image or running multiple processes; that breaks horizontal scaling.
Explain a model registry's purpose and what to store per version
Tests if you treat the registry as a governance bridge between training and production, not just storage. Strong answers cite versioned artifacts, lineage, metrics, dependencies, and approval gates. Red flag: calling it a file dump or experiment tracker.
Expose a trained model as a simple web service
Practical MLOps knowledge from model serialization to serving. Package the model into a standard format, containerize it, expose a REST endpoint behind a load balancer, and add monitoring. A bare Flask server without containers or health checks is a red flag.
Explain model quantization, its benefits, drawbacks, and validation approach
Tests precision trade-offs in production. Answer: define lowering weights from fp32 to int8/int4; cite memory and latency gains versus accuracy loss; validate with downstream benchmarks and shadow A/B. Red flag: treating as lossless or skipping task metrics.
Deploy a trained model as a containerized REST API
This tests practical MLOps fluency. A strong answer covers loading the artifact, wrapping it in a web server, building a Dockerfile, and exposing health and predict endpoints. Red flag: conflating training with serving or omitting resource limits.
What is a model registry's purpose in CI/CD4ML and its CI/CD interaction?
Tests if you see the model registry as the bridge between experiments and production, not just storage. A strong answer explains how CI publishes validated artifacts and CD consumes versioned models. Red flag: calling it a passive file dump without versioning.
What automated tests belong in CI before deploying a classification model?
Name data schema checks, performance regression vs baseline, bias audits, and artifact integrity. Distinguishing code tests from ML-specific CI validation. Only testing the inference API while ignoring model behavior.
Design a CI/CD pipeline that automates model promotion from Staging to Production
Tests whether you treat model promotion as a gated software delivery workflow. Strong answers use registry state-change triggers, automated drift and performance checks, canary deployment gates, and rollback.
What are the primary differences between traditional DevOps and MLOps?
Tests if you know MLOps extends CI/CD to data and probabilistic assets. Contrast code artifacts with models and datasets. Distinguish commit triggers from drift or retraining triggers. Red flag: treating models as static binaries and ignoring data lineage.
MLflow Models Standardize Deployment Packaging
MLflow Models wrap artifacts into a standard package so one pipeline serves sklearn or PyTorch without new deployment code. Teams ship experiments to REST endpoints without Dockerfiles per model. Missing dependency logging lets model load but fail to predict.
Model Registry: Source of Truth for Deployed Models
A model registry is the source of truth for which trained model runs where, turning anonymous artifact files into versioned, staged assets. It matters when you deploy multiple models or need instant rollbacks.
Model Server: The MLOps Deployment Bridge
A model server bridges ML training and production, operationalizing models within your release cycle. Use it when models must become first-class CI/CD citizens. The footgun is treating deployment as a one-time handoff rather than repeatable infrastructure.
Shadow Deployment: Test Models on Real Traffic
Shadow deployment runs a new model on real traffic without serving its predictions, letting you catch data drift before users are affected. It is the safest production validation method, but teams often forget to monitor its latency and resource costs.
ML Model Registry: Source of Truth for Production Models
A model registry is version control for trained models, not just code. It tracks which artifact is running in production, who approved it, and how it was built. Skip it and you get untracked files in S3 with no way to reproduce a production model.
ML Pipeline: Systematic Model Delivery
A machine learning pipeline is the systematic workflow that carries models from data labeling through deployment inside MLOps. It keeps the AI lifecycle repeatable rather than ad hoc. The footgun is treating a one-off notebook as a production pipeline.
How do you add a basic post-deployment health check in CI/CD?
Tests deployment validation beyond exit-code success. Outline: add a post-deploy stage that probes an HTTP endpoint, checks status code and latency, validates critical dependencies, and triggers rollback on failure.
Walk me through deploying a new version using a GitOps workflow
Merge updates manifests; the GitOps controller compares desired state in Git to live state and reconciles. Separation of CI from CD and Git as the single source of truth. Mixing CI with CD or recommending manual deploy.
How would you use distributed tracing to debug a deployment latency issue?
This tests causal request-path analysis beyond aggregate metrics. A strong answer filters traces by the new version, finds the exact regressed span, and compares it to a pre-deployment baseline.
Design an automated rollback process when deployment error rates spike
Gate on error-rate and latency thresholds; use blue-green deploys to limit blast radius; require human approval for stateful rollbacks. judgment when automating rollbacks. instant full rollback from one noisy metric.
What are the four Golden Signals for service health monitoring?
Tests whether you can name the four essential metrics—latency, traffic, errors, saturation—and explain why each matters for detecting user-facing regressions after a release, rather than drowning in infrastructure noise.
Blue/green deployment fails during switch-over with partial decommissioning; recovery and process changes?
Tests whether you can recover when a blue/green rollback path is compromised. Strong answers stop the bleed, revive blue if possible, and mandate keeping blue fully warm until green is stable. Red flag: "just roll back" ignoring partial decommissioning.
What are the major challenges of blue/green deployments with relational databases?
Tests decoupling schema and code changes in stateful blue/green deployments. Strong answers cover the additive-then-deletive pattern, backward compatibility for both app versions, and shared DB risks. Red flag: split DBs with no rollback or sync plan.
What problem can a breaking API change cause during a rolling update?
Tests if you know rolling updates run mixed versions, so breaking API changes crash cross-traffic. Good answer: note old and new pods serve together, watch probes fail, and monitor 5xx spikes. Red flag: claiming Kubernetes isolates versions during rollout.
Get Deployment bites daily.
Five a day, five minutes, offline. With quizzes so it sticks.
Open testing — you’ll join as an early tester.