Loading your roadmap…
Five skill pillars, a sequence of certifications, and the projects that prove them. Track what you're doing now — and what's left.
Overall progress
0%
0 of 31
milestones complete
0
In progress
0/5
Pillars complete
0/6
Certifications
0/5
Projects shipped
You're viewing a read-only preview
Sign in with GitHub to track your own progress — it syncs across every device.
Five framework-agnostic capabilities. Each ring fills as you complete that pillar's checkpoints.
The bedrock — talking to models well and cheaply.
0 / 4 checkpoints done
The most in-demand single pattern in job postings.
0 / 4 checkpoints done
Wiring reasoning, tools, and state into reliable agents.
0 / 4 checkpoints done
The emerging backbone: MCP for tools, A2A for agents.
0 / 4 checkpoints done
The scarce, premium-paying skill — invest heavily here.
0 / 4 checkpoints done
An ordered path — earn them in sequence; each one forces you to consolidate a pillar.
LangChain Academy + Anthropic Academy.
The on-ramp and exam prep — start here.
MCP + agentic architecture; 27% of the exam is agents.
Free for early adopters — time-limited. Do this first.
Foundational AI fluency on your existing AWS base.
Optional quick win — light but cheap and fast.
Bedrock AgentCore, RAG, production generative AI.
The production heavyweight; sit it after Bedrock hands-on.
Multi-agent design, reasoning, orchestration at depth.
Most specialized agentic cert — earn it after shipping agents.
University cohort with live faculty and hands-on projects.
Worth it if you want structure and a university credential.
One project per pillar. These matter more than the certs — ship them to GitHub.
Retrieval app over a real corpus + Ragas/DeepEval harness + reranking.
Tool-using agent with human-in-the-loop and persistence/resume.
Expose a real tool or data source over MCP and consume it.
2–3 coordinating agents; bonus: wire an A2A interaction.
Instrument the projects above with Langfuse + LLM-as-judge + a cost dashboard.