Data Scientist Roadmap
You want to answer hard questions with statistics and models
Full mathematical and statistical foundation, then machine learning end to end.
The route
- Week 1 - Foundations: Strong math fundamentals are non-negotiable for rigorous data science
- Week 2 - Python for ML: Your daily toolkit - NumPy, pandas, matplotlib, scikit-learn
- Week 2 - SQL for Data: Daily tool - extract training data, build features, explore datasets at database scale
- Week 3 - Git & Version Control: Version your notebooks, collaborate on shared codebases, and ensure experiment reproducibility
- Week 3 - Jupyter & Colab: Your primary workspace for exploration, analysis, and presentation
- Week 3 - Probability Theory: Bayes theorem, distributions, expected value - asked in every DS interview
- Week 4 - Statistics: Core interview topic - hypothesis testing, confidence intervals, distributions
- Week 5 - Linear Algebra: PCA, SVD, embeddings - the math behind dimensionality reduction
- Week 6 - Calculus: Understand optimization deeply - gradients, chain rule, integrals
- Week 7 - Optimization: How every ML model learns - gradient descent, loss surfaces, convergence
- Week 8 - ML Algorithms: The core of data science - regression, trees, SVMs, ensembles, evaluation
- Week 9 - Deep Learning: Neural networks and transformers - increasingly required for DS roles
- Week 9 - Computer Vision: Scope a vision request before it becomes a labelling bill - which task, what it costs to annotate, and which metric can score it
- Week 10 - LLM: Modern data science uses LLMs for analysis, summarization, and augmentation
- Week 11 - Agentic AI & Multi-Agent Systems: Modern data science uses AI agents for automated analysis and data workflows
- Week 11 - ML Pipeline: End-to-end workflows - EDA, preprocessing, training, evaluation, deployment
- Week 12 - Information Theory: Entropy, KL divergence, mutual information - understand model decisions deeply
- Week 13 - AI Tools: AI coding tools speed up data exploration, notebook workflows, and model experimentation
- Week 13 - AI on Cloud: Run notebooks and managed ML/GenAI services in the cloud - Vertex AI, Bedrock, Azure ML basics
- Week 14 - Interview Q&A: Scenario-based interview prep - production problems that AI companies actually ask
- Week 14 - AI System Design: The round where a data scientist is asked to ship rather than to analyse. Sizing, serving and evaluation, in the vocabulary the interviewer actually uses