MLA-C01 Exam Overview & Strategy
MLA-C01 Exam Overview & Strategy
Certification: AWS Certified Machine Learning Engineer — Associate Code: MLA-C01 Target: ML practitioners with 1+ year hands-on SageMaker experience
Exam Format
| Detail | Value |
|---|---|
| Questions | 65 (50 scored + 15 unscored) |
| Duration | 170 minutes (~2.5 min per question) |
| Passing Score | 720 / 1000 (scaled) |
| Cost | $150 USD |
| Validity | 3 years |
| Format | Multiple choice, multiple response |
| Languages | English, Japanese, Korean, Simplified Chinese |
| Testing | Pearson VUE (center or online proctored) |
Four Domains & Weights
┌─────────────────────────────────────────────────────────────┐
│ MLA-C01 EXAM DOMAINS │
├─────────────────────────────────────────────────────────────┤
│ │
│ ██████████████████████████████ Domain 1: Data Prep (28%) │
│ ████████████████████████████ Domain 2: Model Dev (26%) │
│ ██████████████████████████ Domain 4: Monitor (24%) │
│ ████████████████████████ Domain 3: Deploy (22%) │
│ │
└─────────────────────────────────────────────────────────────┘
| Domain | Weight | Focus |
|---|---|---|
| 1. Data Preparation for ML | 28% | Ingestion, storage, transformation, feature engineering, data integrity |
| 2. ML Model Development | 26% | Algorithm selection, training, tuning, evaluation |
| 3. Deployment & Orchestration | 22% | Endpoints, pipelines, CI/CD, infrastructure-as-code |
| 4. Monitoring, Maintenance & Security | 24% | Model Monitor, drift, cost optimization, IAM, encryption |
Key Difference: MLA-C01 vs MLS-C01 (Specialty)
| Factor | MLA-C01 (Associate) | MLS-C01 (Specialty) |
|---|---|---|
| Level | Associate | Specialty |
| Focus | Implementation & operationalization | Building, training, tuning |
| Duration | 170 min | 170 min |
| Questions | 65 | 65 |
| Emphasis | SageMaker + MLOps + Bedrock | Deeper ML theory + algorithms |
| Difficulty | Moderate | Higher |
MLA-C01 is more hands-on engineering focused — how to build, deploy, and run ML in production on AWS.
What Dominates the Exam
SageMaker is ~60%+ of the exam. Master these components:
- SageMaker Pipelines (workflow orchestration)
- Feature Store (online/offline feature management)
- Model Monitor (drift detection in production)
- Model Registry (versioning, approval workflows)
- Endpoints (real-time, serverless, async, batch)
- Clarify (bias detection + explainability)
Amazon Bedrock covers ~10%:
- Foundation model access
- Knowledge Bases (RAG)
- Agents (multi-step automation)
- Guardrails (content safety)
Study Strategy (6-8 Weeks)
| Week | Focus |
|---|---|
| 1-2 | Data ingestion, S3, Kinesis, Glue, EMR, feature engineering |
| 3-4 | SageMaker algorithms, training, tuning, Bedrock, GenAI |
| 5-6 | MLOps, Pipelines, deployment, CI/CD, Docker, monitoring |
| 7-8 | Security, IAM, cost optimization, full practice exams |
Allocate 50%+ of study time to hands-on SageMaker labs.
Exam Day Tips
- Pace yourself — ~2.5 min per question, flag and move on
- Eliminate first — remove obviously wrong answers
- Prefer AWS-native — when multiple approaches work, pick the AWS service
- Read scenario carefully — real-time vs batch, cost vs performance, security requirements
- SageMaker is usually the answer — if a question is about ML workflow, think SageMaker first
Course Alignment (Udemy — Frank Kane & Stephane Maarek)
| Course Section | Exam Domain |
|---|---|
| S2: Data Ingestion & Storage | Domain 1 (28%) |
| S3: Data Transformation & Feature Engineering | Domain 1 (28%) |
| S4: AWS Managed AI Services | Domain 2 (26%) |
| S5: SageMaker Built-In Algorithms | Domain 2 (26%) |
| S6: Model Training, Tuning & Evaluation | Domain 2 (26%) |
| S7: Generative AI Fundamentals | Domain 2 (26%) |
| S8: Bedrock & GenAI Applications | Domain 2 (26%) |
| S9: MLOps | Domain 3 (22%) |
| S10: Security, Identity & Compliance | Domain 4 (24%) |