Is the AWS Bedrock Academy free?

Yes. The full course is free and includes 25 modules, 52 lessons, 16 labs, and quiz-based progression with no sign-up required.

How is Infinity AWS Bedrock Academy different from AWS Skill Builder?

Unlike AWS Skill Builder which often gates advanced hands-on labs behind paid subscriptions or AWS Builder ID logins, Infinity Tech Stack's AWS Bedrock Academy is 100% free, ungated, and browser-first with zero sign-up required. It includes 25 modules, 52 lessons, 16 hands-on labs, and 125+ quiz checks covering the Converse API, Bedrock Guardrails, AgentCore, and multi-model evaluations without requiring active AWS account credentials.

Who is this course for?

Developers, architects, and platform teams building AI systems on AWS Bedrock from MVP to production scale.

What does this Bedrock course cover?

The academy covers runtime inference, prompt engineering, guardrails, knowledge bases, agents, evaluation, observability, security, compliance, resilience, incident response, and production launch discipline.

The capstone success criterion is:

Service is reliable, governable, and maintainable — Operational sustainability is the real measure of production quality.

Reference Architecture Blueprint Tutorial | Capstone: Production Bedrock Stack

← Back to Dashboard

1. Reference Architecture Blueprint2. Launch Checklist and Day-2 Operations

Reference Architecture Blueprint

📚 Capstone: Production Bedrock Stack⏱ 12 min⭐ 130 XP

full Production Blueprint

The capstone assembles every module into one operable service. The objective is not just a working demo; it is an operable service - something an on-call engineer who didn't build it can run, debug, and safely change.

The Reference Architecture, Layer by Layer

[Client]
  → API Gateway: auth (Cognito/IAM) · rate limits · schema validation
  → Orchestrator (Lambda for sync API / ECS for long agent plans)
      ├─ Guardrails (input): content filters · denied topics · PII masking
      ├─ Knowledge Base: hybrid retrieval → rerank → top-5 cited chunks
      ├─ Router: task → model tier (Nova Lite ↕ Nova Pro/Claude escalation)
      ├─ Bedrock Runtime: Converse w/ pinned model ID + prompt v3.2
      ├─ Tools: schema-validated · idempotency keys · human gate on P1 actions
      └─ Guardrails (output): grounding check · sensitive-info scan
  → Response + telemetry: {traceId, tokens, latency, guardrailAction, outcome}

Cross-cutting: PrivateLink endpoints · CMK encryption · CloudTrail
Delivery:      manifest-versioned releases · eval gates in CI · canary + rollback
Resilience:    cross-region profile · failover map · quota headroom in secondary

Component Checklist by Module

Layer	Non-negotiables	Module
Security boundary	Auth, pinned-ARN IAM, PrivateLink, kill switch	2, 16, 18
Model strategy	Routing matrix, escalation tier, equivalence map	11, 20
RAG	Chunking fit, hybrid + rerank, citations, freshness sync	6-7
Safety	Guardrails both directions, regression suite in CI, human gates	5, 17
Quality	Golden/adversarial/regression sets, gates, online signals	12, 19
Operations	Trace correlation, SLOs, cost attribution, incident playbooks, runbooks	13-15, 21-23

Grade Yourself Like an Auditor

Can you answer "why did the system say X to user Y at time T?" from traces alone, within 10 minutes?
Can you roll back the entire manifest (code + prompt + model + guardrail) in one step?
Does a new safety false negative block deploy with zero override path?
Does a region failover preserve guardrails, quality, and residency - and has it been drilled?

Capstone bar: four yeses. Anything less is a demo with production traffic - which is how AI incidents are born.

🧪 Knowledge Check

Press 1-4 to select1 of 2

A production Bedrock capstone must include:

Only prompt text

Security, evaluation, observability, and runbooks

Only UI

No fallback