Week 01
4 playableFDE Mindset & Toolkit
Linux diagnostics, Git plumbing, and the discovery question that replaces a missing spec.
90-day track
Twelve weeks,
one map
All 61 lessons across the 12 weeks are playable in the beta. Pick up where the map shows you left off.
- Foundation
- Enterprise AI
- Field delivery
Foundation
Weeks 1–4Week 02
6 playableProduction Python
Type hints, structured logs, retries, and circuit breakers that fail fast inside a customer VPC.
Week 03
5 playableData audits & API trust boundary
Schema forensics, current-row SQL, Pydantic at the edge, Redis in front of Postgres, and dbt tests that halt bad data.
Week 04
5 playableCloud & containers
Multi-stage images, private-by-default networking, workload identity, a secrets manager, and a SecOps one-pager before the review.
Enterprise AI
Weeks 5–8Week 05
5 playableLLM gateways
Token cost, latency budgets, hard ceilings, JSON schemas, and a fallback provider so one outage does not stop Helix.
Week 06
6 playableRAG and evals
Retrieve instead of pasting PDFs, keep tables intact, hybrid search with RRF, and ACL before retrieval.
Week 07
5 playableAgents
Supervisor graphs, error feedback, step caps, human approval for writes, and trimmed memory.
Week 08
5 playableMCP & identity
MCP as the door to customer systems, tools vs resources, OAuth/SAML/SCIM, act-as-user, least privilege.
Field delivery
Weeks 9–12Week 09
5 playableKubernetes & Terraform
Plan before apply, modular Terraform, perimeters with dry-run, and deploys only through CI.
Week 10
5 playableAI security & red teaming
Treat docs as data, mask PII, decode 403s, red-team the bot, and pass the CISO questionnaire.
Week 11
5 playableObservability & break-glass
Trace IDs, token alerts, golden sets, model A/B, and incident updates that lead with impact.
Week 12
5 playableDelivery & handover
SOW non-goals, a renewal number, a runbook, three-deltas productization, and the capstone defense.