hub.gazar.dev · course

From Senior to Staff

Live session decks and project briefs for the Maven cohort. Twelve 90-minute sessions plus six projects, each slotted into the week it belongs. maven.com/gazar/senior-to-staff-engineer

Week 1 · The 6 Core Trade-offs & Estimation

Jun 2 to Jun 7

Week 2 · Data Stores & Data at Scale

Jun 8 to Jun 14

31 slides

S3 · Relational vs Document vs Key-Value

Picking a store from access patterns, not familiarity. Wed 6/10, 90 min.

/senior-to-staff/s03-data-stores
cheat sheet Reference · cheat sheet

Data Stores Cheat Sheet

The four questions, six families side by side, a card per database (Postgres, Mongo, Redis, Cassandra, …), ERD design and normal forms. Companion to S3.

/cheatsheets/data-stores
23 slides

S4 · Data at Scale: Sharding, Replication, Caching

What changes when one node stops being enough. Fri 6/12, 90 min.

/senior-to-staff/s04-data-at-scale
cheat sheet Reference · cheat sheet

Data at Scale Cheat Sheet

How to scale each database (Postgres, MySQL, Mongo, Redis, Cassandra, DynamoDB, …) for reads vs writes, the lever you pull, and the trade-off. Sharding, replication and caching side by side. Companion to S4.

/cheatsheets/data-at-scale
cheat sheet Reference · cheat sheet

PostgreSQL vs MongoDB Cheat Sheet

SQL vs NoSQL side by side: the two data shapes, how each handles ACID, what only one can do, when to choose which, and interview questions with answers. Companion to S4.

/cheatsheets/postgres-vs-mongodb
cheat sheet Reference · cheat sheet

Caching Strategies Cheat Sheet

The first lever for scaling reads, in full: the hit-ratio math, read strategies (cache-aside vs read-through) and write strategies (write-through, write-back, write-around) with sequence diagrams, how they pair, consistency and the stale window, eviction (LRU/LFU/TTL), invalidation and the delete-vs-update trap, the failure modes (stampede, hot key, penetration, avalanche) and their fixes, the layers from browser to CDN to Redis, CDN delivery, which technology to use (Redis vs Memcached and friends), cache architecture and topology (local vs distributed vs near-cache, sharding, consistent hashing), key design and metrics, with a decision tree and interview Q&A. Companion to S4.

/cheatsheets/caching-strategies
1-page brief Project 2 · due Jun 19

P2 · ADR #1: Database Decision

Record a data-store choice as a one-page ADR, with the trade-off named.

/senior-to-staff/p2-adr-database

Week 3 · APIs & Distributed Systems

Jun 15 to Jun 21

30 slides

S5 · API Design & Networking Fundamentals

Contracts, versioning, and the network you can't ignore. Wed 6/17, 90 min.

/senior-to-staff/s05-api-design
cheat sheet Reference · cheat sheet

API Design & Networking Cheat Sheet

The contract and the network on one page: REST vs gRPC vs GraphQL and when to choose which, contract design (idempotency, pagination, versioning, status codes and error shapes, rate limiting), HTTP/1.1 vs 2 vs 3, the DNS/TCP/TLS handshakes, keep-alive and connection pooling, L4 vs L7 load balancing and the CDN edge, a decision table and interview Q&A. Companion to S5.

/cheatsheets/api-design
cheat sheet Reference · cheat sheet

Security & Auth Cheat Sheet

Everything security and identity on one page, with the flows drawn out. Authn vs authz; how identity is carried (sessions, JWTs, opaque tokens); OAuth 2.1 authorization-code + PKCE, OpenID Connect, SAML SSO and federation, passkeys/WebAuthn, the BFF token-handler pattern, service-to-service mTLS + SPIFFE + token exchange (RFC 8693); where to enforce auth; the authorization models (RBAC, ABAC, ReBAC/Zanzibar); plus password storage (argon2 + salt), encryption in transit and at rest, and the common web attacks (injection, XSS, CSRF, SSRF, IDOR) and their defenses. Decision guide and interview Q&A. Companion to S5.

/cheatsheets/security-auth
28 slides

S6 · Distributed Systems Patterns

Consensus, idempotency, sagas, and the failures behind them. Fri 6/19, 90 min.

/senior-to-staff/s06-distributed-patterns
cheat sheet Reference · cheat sheet

Backend Clocks & Time Cheat Sheet

The "no global clock" problem unpacked for backend engineers and interviews: physical sync (NTP, PTP), wall vs monotonic (with code), logical clocks (Lamport, vector), hybrids (HLC, TrueTime), time-based IDs (Snowflake, ULID, UUIDv7), backend rules (UTC, JWT skew, fencing tokens), a decision tree, and interview Q&A. Companion to S6.

/cheatsheets/clocks-and-time
cheat sheet Reference · cheat sheet

Consistency & Consensus Cheat Sheet

How distributed systems agree and what it costs: CAP and PACELC, the consistency spectrum (linearizable to eventual), quorums (W+R>N), replication topologies, consensus with Raft and majorities, and cross-service transactions (2PC vs saga), with a decision table and interview Q&A. Companion to S6.

/cheatsheets/consistency-consensus

Week 4 · Architectural Styles & DDD

Jun 22 to Jun 28

30 slides

S7 · Architectural Styles: Monolith to Microservices

The spectrum, and when each point on it is the right call. Wed 6/24, 90 min.

/senior-to-staff/s07-architectural-styles
cheat sheet Reference · cheat sheet

Kafka vs RabbitMQ Cheat Sheet

Messaging for interviews: the log-vs-broker distinction, core concepts for each (partitions, consumer groups, offsets vs exchanges, queues, bindings, acks), delivery semantics, ordering, when to use which, common patterns, and interview Q&A. Companion to S6 and S7.

/cheatsheets/kafka-vs-rabbitmq
30 slides

S8 · Domain-Driven Design for Architects

Bounded contexts and service boundaries that don't leak. Fri 6/26, 90 min.

/senior-to-staff/s08-ddd
1-page brief Project 3 · due Jul 3

P3 · ADR #2: Service Boundary Decision

Decide where one service boundary goes, justified by a real seam, not the org chart.

/senior-to-staff/p3-adr-service-boundary
cheat sheet Reference · cheat sheet

Event Storming for Food Delivery System

A full event-storming session on a real project (a food-delivery platform), with the sticky-note board growing as you scroll: chaotic events, the timeline, hotspots and pivotal events, actors and external systems, commands, aggregates, policies, read models, until the bounded contexts emerge and become services. Companion to S8.

/system-designs/event-storming-food-delivery-system
case study Case study · real app

Event Storming, for Club Control Panel

ClubCP, a real club-management system, taken through a full event-storming session, the same steps, real answers: every sticky traced from the actual code as the board grows down the page. The timeline of a member attending a paid event, the real hotspots (dual payment providers, idempotent capture, capacity & waitlist, no-show emails, bounce suppression), then actors, commands, aggregates, policies and read models, until ClubCP's bounded contexts emerge. The whole workshop on real stakes.

/system-designs/event-storming-club-control-panel

Week 5 · Reliability & Operability

Jun 29 to Jul 5

17 slides

S9 · Designing for Failure

Blast radius, fail-open vs fail-closed, graceful degradation. Wed 7/1, 90 min.

/senior-to-staff/s09-designing-for-failure
cheat sheet Reference · cheat sheet

Resilience Patterns Cheat Sheet

The whole failure-handling toolkit on one page: the four bounding patterns (timeout, retry with backoff and jitter, circuit breaker, bulkhead), fallback and graceful degradation, load shedding, correctness under retry (idempotency, at-least-once vs exactly-once, dead-letter queues), fail-open vs fail-closed per path, operational glue, a symptom-to-pattern decision table, and interview Q&A. Companion to S9.

/cheatsheets/resilience
18 slides

S10 · Observability & Operational Readiness

SLIs/SLOs, the three pillars, and what "ready to run" means. Fri 7/3, 90 min.

/senior-to-staff/s10-observability
cheat sheet Reference · cheat sheet

Observability & SLO Cheat Sheet

Reliability as a number on one page: SLI / SLO / error budget (with what each nine buys in minutes), the three pillars (metrics, logs, traces) and their pitfalls, the RED and USE methods, symptom and burn-rate alerting, an operational-readiness checklist, a question-to-pillar decision table, and interview Q&A. Companion to S10.

/cheatsheets/observability
1-page brief Project 4 · due Jul 6

P4 · ADR #3: Capstone Decision

Your hardest call, reasoned end to end: number, trade-off, failure, operability.

/senior-to-staff/p4-adr-capstone
1-page brief Project 5 · due Jul 6

P5 · System Design Document

The whole capstone system on a few pages. The artifact that most looks like the job.

/senior-to-staff/p5-design-doc

Week 6 · Design Review

Jul 6 to Jul 9

6 slides

S11 · Live Group Design Session

Designing together under pushback. Tue 7/7, 60 min.

/senior-to-staff/s11-group-design
6 slides

S12 · Design Review Presentations

Present, defend, and read the room. Thu 7/9, 90 min.

/senior-to-staff/s12-design-review
1-page brief Project 6 · capstone

P6 · Design Review Presentation

Present and defend your design live. Rehearsed in S11, presented Thu 7/9.

/senior-to-staff/p6-design-review
worked design Reference · worked system design

Design Twitter / X: the home timeline

The classic prompt, worked end to end with the course method: requirements, estimation, API, data model, a high-level architecture diagram, the fan-out deep dives (write vs read, the celebrity hot-key problem, ranking), reliability (S9) and observability (S10), and the trade-offs along the six core axes. Applies the whole course to one interview question.

/system-designs/twitter
worked design Reference · worked system design

Design WhatsApp: chat & messaging

Realtime messaging worked end to end: requirements, estimation (connections as the deciding number), the persistent-socket protocol, data model, a high-level architecture diagram, the deep dives (holding hundreds of millions of connections, delivery and ordering semantics, store-and-forward for offline), reliability (S9) and observability (S10), and the trade-offs along the six core axes.

/system-designs/whatsapp
worked design Reference · worked system design

Design Stripe: payments

The correctness-first design, worked end to end: requirements, estimation (why QPS is not the constraint), the idempotent API, data model, a high-level architecture diagram, the deep dives (idempotency keys, a double-entry ledger and exactly-once accounting, reconciliation and fail-closed by default), reliability (S9) and observability (S10), and the trade-offs along the six core axes. The deliberate counterpoint to Twitter and WhatsApp.

/system-designs/stripe

More worked system designs

Reference · the method, then the classic interview prompts worked end to end

the method Reference · the method on one page

The System Design Interview Framework

The whole method on one page: how to run the hour and the seven steps (requirements and scope, estimation, API, data and stores, high-level design, deep dives, trade-offs), plus what actually scores, time budgets, and the common traps. The umbrella the eight worked designs below all apply. Bring it to any "design X" prompt.

/system-designs/framework
worked design Reference · worked system design

Design Uber: ride-sharing

The geospatial design, worked end to end: requirements, estimation (a million location writes per second as the deciding number), API, data model, a high-level architecture diagram, the deep dives (geospatial indexing with cells / geohash / S2 / H3, the real-time location firehose, matching and dispatch without double-booking), reliability (S9) and observability (S10), and the trade-offs along the six core axes.

/system-designs/uber
worked design Reference · worked system design

Design YouTube: video streaming

The video design, worked end to end: requirements, estimation (egress bandwidth as the deciding constraint), API, data model, a high-level architecture diagram, the deep dives (the async upload + transcode pipeline and rendition ladder, CDN delivery with adaptive bitrate, and read scale across the long tail), reliability (S9) and observability (S10, rebuffer ratio), and the trade-offs along the six core axes.

/system-designs/youtube
worked design Reference · worked system design

Design Dropbox: file sync

The file-sync design, worked end to end: requirements, estimation (edit-size to file-size as the deciding ratio), a content-addressed chunk API, data model, a high-level architecture diagram, the deep dives (content-defined chunking and dedup, the cursor-based sync protocol, and conflict resolution via conflicted copies), reliability (S9) and observability (S10, dedup ratio), and the trade-offs along the six core axes.

/system-designs/dropbox
worked design Reference · worked system design

Design Google Search: web search

The search design, worked end to end: requirements, estimation (the index does not fit on one machine, so scatter-gather), API, data model, a high-level architecture diagram, the deep dives (the crawler frontier, the inverted index with document-partitioned scatter-gather, two-phase ranking and typeahead), reliability (S9, drop slow shards on a deadline) and observability (S10, latency + completeness), and the trade-offs along the six core axes.

/system-designs/google-search
worked design Reference · worked system design

Design Ticketmaster: event ticketing

The high-contention design, worked end to end: requirements, estimation (buyers-to-seats ratio at onsale as the deciding number), a hold-then-purchase API, data model, a high-level architecture diagram, the deep dives (high-contention inventory with no double-booking via an atomic conditional update, the onsale spike and virtual waiting room, seat holds and expiry without a race), reliability (S9, fail closed on the seat) and observability (S10, double-booked = 0), and the trade-offs along the six core axes.

/system-designs/ticketmaster

Each project page explains the whole brief in detail, with a fill-in template, the rubric, and a collapsible instructor talk-track. Bonus (async): The Least-Wrong Decision Model · The Architecture Communication Ladder · Interview Mode: System Design Rounds.