Data Science learning roadmap

What you will learn

1. 01

   ML Fundamentals

   Frame a prediction problem, define its target and cost of error, and build a reproducible baseline before tuning models.

   Problem & target framing · Baselines · Reproducible pipelines

   14h
2. 02

   Feature Engineering

   Create leakage-safe feature transformations and prove that each retained feature improves a cross-validated baseline.

   Encodings & scaling · Leakage traps · Feature selection

   10h
3. 03

   Model Building & Evaluation

   Evaluate errors with business-aligned metrics, calibration, subgroup analysis, and an untouched final test set.

   Metrics & thresholds · Calibration · Error & subgroup analysis

   12h
4. 04

   Experimentation & A/B Testing

   Design a powered experiment with a decision rule, guardrail metrics, validity checks, and an honest interpretation of uncertainty.

   Experiment design · Power & effect size · Validity & decision rules

   12h
5. 05

   Model Deployment

   Package the chosen model behind a tested API with versioned artifacts, input validation, latency measurement, and rollback instructions.

   Validated inference API · Artifact versioning · Latency & rollback

   12h
6. 06

   Advanced Modeling

   Choose an advanced method that matches the data structure, validate it correctly, and ship it only when measured value justifies added cost and risk.

   Problem-shaped model choice · Grouped & temporal validation · Complexity trade-offs

   18h
7. 07

   Responsible Production Capstone

   Ship the cumulative project with reproducible training, a model card, monitored inference, responsible-use checks, and a stakeholder decision memo.

   Model documentation · Monitoring contract · Responsible release

   16h