Curriculum
Building an AI-Native, Decentralized Social Ecosystem A 90-Lesson Premium Course for Engineers, Architects & Technical Leaders
What You Can Build After This Course
Completing all 90 lessons, you can:
Replace six separate infrastructure services with four purpose-fit, open-source alternatives and defend that decision with benchmarks
Build systems that understand the meaning of content, not just its keywords
Implement cryptographic identity and content provenance from scratch, without third-party libraries doing the hard work for you
Design federated social platforms that interoperate with the global fediverse
Write systems where users provably control their own data, identity, and algorithms
Deploy infrastructure that predicts its own failures and heals without human intervention
Profile any production system at the kernel level without touching application code
Design zero-knowledge proof circuits for privacy-preserving verification
1 | The Polyglot Tax — Why Your Stack Is Costing You More Than You Think
2 | Edges as Citizens — Modeling the Social Graph Natively
3 | The Traversal Query — Fetching a Timeline Through Relationships
4 | The Living Database — Real-Time Subscriptions Without a Message Queue
5 | The Honest Benchmark — SurrealDB vs the Stack It Replaces
6 | Killing the JVM — Setting Up Redpanda
7 | The Append-Only Truth — Designing for 100% Data Durability
8 | The Undo Button — Point-in-Time Recovery via Event Replay
9 | Commands vs Events — Designing the Write Path
10 | Projections — Building Pre-Computed Read Models
11 | The Consistency Window — Handling Eventual Updates Gracefully
12 | Stepping Through Time — Building the Event Debug Interface
13 | Meaning Over Keywords — Installing and Tuning Qdrant
14 | Text to Numbers — Generating Embeddings on CPU
15 | The Meaning Index — Semantic Search over 100K Tweets
16 | The Precision Layer — Hybrid Search with Metadata Filters
17 | The Object Store — Setting Up SeaweedFS
18 | The Data Lifecycle — Automated Hot/Warm/Cold Migration
19 | Shrinking the Index — Vector Quantization in Practice
20 | The Sync Pipeline — Change Data Capture from SurrealDB to Qdrant
21 | The Unified API — One Interface, Four Optimal Stores
22 | The Local Runtime — Serving LLMs Without a Cloud Bill
23 | The Model Selection Matrix — Phi-3 vs Gemma vs Llama 3.2 on CPU
24 | Tokens in Real Time — Building Streaming AI Interfaces
25 | The Classification Pipeline — Tagging the Redpanda Stream at Scale
26 | Beyond Keywords — Context-Aware Toxicity Detection
27 | Taming the Output — Forcing LLMs to Return Structured JSON
28 | The Backpressure Problem — Handling AI Lag During Traffic Spikes
29 | The Memory Machine — Building RAG Retrieval
30 | The Grounded Answer — RAG Synthesis with Source Attribution
31 | The Rolling Summary — Trending TL;DRs via Sliding Windows
32 | Delta Summarization — Updating Summaries Without Reprocessing Everything
33 | The Persona Library — Versioned System Prompts for Social AI
34 | The Explanation Engine — Chain-of-Thought for Moderation Appeals
35 | The Tool-Using Agent — ReAct with Search
36 | The Autonomous Manager — Agents That Plan and Execute
37 | The Guardrail Layer — Human-in-the-Loop for Agentic Actions
38 | The Interest Vector — Semantic Personalization
39 | The Living Profile — Interest Vectors That Evolve
40 | The Model Router — Routing Tasks to the Right Inference Engine
41 | The Inference Cache — Eliminating Redundant Computation
42 | The Feedback Loop — Prompts That Learn from User Behavior
43 | The Multilingual Pipeline — Language-Agnostic Classification
44 | The Explanation UI — Making AI Decisions Transparent
45 | Phase 2 Milestone — The Autonomous Community Manager
46 | Keys Instead of Passwords — Implementing DIDs
47 | Key Hygiene — Rotation, Recovery, and Device Management
48 | The Unforgeable Post — Ed25519 Signing for Every Tweet
49 | The Provenance Chain — Tracing Authorship Through Retweets
50 | Metadata Integrity — Signing Hashtags and Timestamps
51 | The Trust Badge — Verifiable Credentials for Account Verification
52 | The Issuer Service — Building the Verification Authority
53 | Verifying Offline — No Issuer Required
54 | The Authentic Media Pipeline — C2PA Content Provenance
55 | The Deepfake Detector — AI-Generated Image Classification at Upload
56 | The Trust Layer UI — Rendering Provenance in the Feed
57 | Opening the Federation — ActivityPub Actor Discovery
58 | The Inbox — Receiving Posts from Mastodon
59 | The Outbox — Broadcasting NEXUS Posts to the Fediverse
60 | The Privacy Proof — ZK Age Verification
61 | The Anonymous Poll — One Vote Per Person, Zero Linkage
62 | Noise as Protection — Differential Privacy for Trending Data
63 | ε-DP for User Profiles — Hiding Individual Interests
64 | The Personal Data Store — Local-First User Vaults
65 | The Capability Model — Scoped, Revocable Data Access
66 | Phase 3 Milestone — The Zero-Auth Social Network
67 | The Three Continents — Simulating a Multi-Region Edge on One Laptop
68 | The Smart Router — Geo-Aware Traffic Direction
69 | Stale is Fast — Edge Caching with Stale-While-Revalidate
70 | The Cold Start Problem — Proactive Cache Warming
71 | The Plugin Host — Installing the Wasmer Runtime
72 | Feed Logic as a Plugin — User-Defined Ranking in WASM
73 | The Sandbox — Hard Resource Limits for Untrusted Code
74 | Rust to WASM — Compiling a High-Performance Ranker
75 | Hot Swap — Updating Algorithms Without Restarting
76 | The Plugin Marketplace — Signed Distribution of Community Algorithms
77 | Offline by Default — CRDT-Based Sync Architecture
78 | The Math of Consistency — Proving CRDTs Are Correct
79 | Personalization at the Edge — Response Variants Without Origin Calls
80 | Rules at the Edge — Decentralized Geo-Blocking
81 | Phase 4 Milestone — “Bring Your Own Algorithm” Feed
82 | Deliberately Breaking Things — Chaos Engineering with Pumba
83 | The Self-Healing Layer — Automated Service Recovery
84 | The Failure Absorber — Circuit Breakers on Every Service Boundary
85 | The Zero-Overhead Observer — eBPF Profiling with Parca
86 | The Regression Detector — Automated Performance Comparison Across Deploys
87 | The Zero-Trust Mesh — mTLS Between Every Service
88 | The Predictive Scaler — AI-Driven Pre-Scaling
89 | The Graduation Challenge — Your Own Cross-Phase Feature
90 | The Living System — NEXUS 2.0 Boots in One Command
NEXUS 2.0 · System Design Twitter Course · 90-Lesson Premium Course
One lesson per day · Ships working code every day · Every tool free and open source
That’s a massive curriculum you’ve got there. Would you like me to help you summarize any specific Phase or create a study guide for these lessons?