20posts ·all writing
What a forward deployed engineer does on AWS, how the role differs from SA work, the four-layer skill stack, and one project that proves you can do it.
#aws #ai-agents #claude-code #certifications
A reference stack for running AI agents in production on AWS: WAF, identity, Bedrock Guardrails, Cedar tool policies, observability, cost and adoption.
#ai-agents #aws #security #platform-engineering
· via Spacelift
Golden paths as products: a few templates with minimal inputs and policy built in, versioned and owned, which agents use through the same API as humans.
#platform-engineering #iac #ai-agents
· via Spacelift
Agents are a new class of platform user. Give each one an identity, an owner, spend caps and audit logs, then widen its autonomy in stages as it earns trust.
#ai-agents #platform-engineering #security
A map of AI SRE agents in 2026: where they fit in the incident workflow, the main tools by category, guardrails, platform readiness and what to measure.
#ai-agents #devops-tooling #platform-engineering
· via Spacelift
Five layers of guardrails for AI-written infrastructure code: before generation, at review, at apply, after apply, and the policy that ties them together.
#ai-agents #iac #security #platform-engineering
How to build an agent control plane on AWS: lifecycle, identity, tool governance, policy, guardrails, sandboxed execution, and detection with AgentCore.
#ai-agents #aws #platform-engineering
· via Spacelift
A nine-step roadmap for modernizing an aging Terraform estate: state surgery, guardrails, self-service, and why AI enters the workflow last.
#iac #platform-engineering #ai-agents
· via Spacelift
How Claude Code fits IaC work in practice: a mental model, the use cases that pay off first, and the guardrails that keep an agent out of state.
#claude-code #iac #ai-agents
· via Gruntwork
Six defense layers for running coding agents against IaC: credential isolation, permission modes, allow and deny rules, CLAUDE.md, hooks, sandboxing.
#claude-code #ai-agents #iac
· via AWS
An agent-driven migration walkthrough: Kiro CLI with MCP servers discovers an EC2 workload, generates the ECS Express Mode setup and cleans up after.
#aws #ai-agents
· via Spacelift
Running AI agents as production workloads: core deployment principles, Amazon Bedrock AgentCore, and a Terraform walkthrough for shipping an agent.
#ai-agents #iac #aws
· via Spacelift
MCP without the hype: the architecture, how servers and clients talk, a demo project, and where the protocol sits next to plain APIs and tools.
#ai-agents #devops-tooling
· via StackGuardian
Where AI lands in platform work: infrastructure automation, smarter self-service, predictive reliability, security and cost, with adoption stages.
#platform-engineering #ai-agents
· via Qodo
Test coverage beyond the percentage: what the metrics really measure, where AI improves gap analysis and generation, and the risks of trusting it.
#ai-agents #devops-tooling
· via Qodo
TDD and code assistants combined: how tests written first constrain generation, a working AI-TDD loop, and the advantages with their caveats.
#ai-agents #devops-tooling
· via Qodo
A selection checklist for AI code assistants: language support, integration, context handling, output quality, security terms, cost and team fit.
#ai-agents #devops-tooling
· via AWS
The business case for training existing staff on generative AI: timing, a customer example of scaling with a small team and the AWS learning paths.
#aws #ai-agents
· via Spacelift
A practical look at Amazon Bedrock: model access, key features, pricing logic, use cases and a tutorial for getting a first application running.
#aws #ai-agents
· via Spacelift
Amazon CodeWhisperer in practice: setup with a Builder ID, day-to-day usage, and a demo writing Terraform to see where the suggestions help.
#aws #ai-agents