Frontend Labs
⋯

What this curriculum actually is

Despite the folder name, this is not a broad "learn Go in 90 days" plan — it's a single continuous capstone: build a from-scratch cloud-edge security proxy (WAF → DDoS mitigation → bot detection → production hardening → integration), ending in a ~5,000-line single static binary. Weeks 1–6 lay Go foundations specifically in service of that proxy (config/storage, concurrency, context/IO, generics/profiling, TCP, TLS/crypto). Weeks 7–17 build the proxy itself, layer by layer, with weeks 12–17 forming one unbroken integration arc.

This is genuinely rare depth in network security engineering — PKI/mTLS, JA3 fingerprinting, 5 rate-limiting algorithms from scratch, SYN-flood/Slowloris defense, WAF normalization bypass handling, FD-inheritance zero-downtime restarts, W3C trace-context parsing. Most Go courses never go near any of this.

The cost of that depth is breadth. Because every day serves the one proxy, entire categories of "everyday Go SME" work never appear at all, in 120 days:

Never appears, anywhere in weeks 1–17
database/sql, pgx, transactions, migrations — zero persistence
gRPC / protobuf
Message queues (Kafka/NATS/SQS), event-driven/async patterns
Kubernetes as producer — no client-go, controller-runtime, CRDs, operators (only as consumer: liveness/readiness semantics)
CLI tooling (cobra/viper) — despite this being a named target niche
Prometheus (client_golang) or OpenTelemetry SDKs — metrics/tracing are hand-rolled (expvar, custom ring-buffer spans) throughout
GC/memory tuning (GOGC, GOMEMLIMIT), escape analysis, benchstat
go test-based unit tests until Week 12 — days 1–81 are almost entirely func main() demos, including a dedicated -race-auditing day with no _test.go file
Password hashing (bcrypt/argon2) — a striking gap given an entire "API Security" week

Generated from real-world-projects-and-sme-gaps.md · 8 Oct 2026 16:36