Systems & Databases
Mission. Give the data scientist credibility with engineers by understanding what code runs on.
Scope
- SQL depth: window functions, CTEs, query plans
- Database internals: indexes, transactions, isolation
- Caching and when it lies to you
- Networking basics and latency across boundaries
- Designing and consuming APIs
- Distributed-systems basics
Learning objectives
- Read a query plan and fix a slow query
- Choose an index and explain why; reason about transactional guarantees
- Design a small API and reason about latency across a network boundary
Lessons
This track is scaffolded and waiting for content. See the curriculum dashboard for the roadmap and progress across all tracks.
Planned lessons (20)
Wave 2 — breadth (provisional roadmap) — 20 lessons
databases
- How a Query Becomes Results Parsing, planning, execution: what the database does with your SQL.
- Joins, CTEs, and Readability SQL another engineer can read six months later.
- Window Functions Ranking, running totals, and per-group computation without self-joins.
- Reading a Query Plan EXPLAIN output, costs, and finding the slow step.
- Indexes and B-Trees How index structure turns scans into lookups — and when it doesn't.
- Composite and Covering Indexes Column order, included columns, and index-only reads.
- Transactions and ACID Atomicity through durability: what a transaction promises.
- Isolation Levels Read phenomena and the tradeoffs between safety and speed.
modeling
- Normalization Normal forms without the ceremony, and when to denormalize.
- Analytics Schemas Star and snowflake: modeling for analysis instead of transactions.
systems
- Caching: Where and Why Layers of caching, hit rates, and what each costs.
- Invalidation and Staleness Expiry, events, and the staleness you can tolerate.
- HTTP and REST Essentials Requests, responses, status codes, and the contract of an API.
- Designing an API for Data Pagination, filtering, payloads: APIs analysts don't hate.
- Latency Is the Enemy Network round trips, batching, and the physics of slow.
- Files, Formats, and Storage Row vs column, compression, and the cost of your choices.
- Containers for Data Science Docker images that make environments portable and boring.
- Threads, Processes, and the GIL What actually runs in parallel in Python, and how to use it.
- Distributed Systems Fallacies The assumptions that break when you add a second machine.
projects
- Case Study: Speeding Up a Slow Pipeline Profile, diagnose, fix: a pipeline from hours to minutes.