Ivan Leo on Gemini interactions, video tools, and managed agents, alongside builders using agentic coding to solve everyday problems.

deepmind london: ivan leo on agents, and the people building with them ivan leo on gemini interactions, video tools, and managed agents, alongside builders using agentic coding to solve everyday problems. ai gcp developer-experience architecture-patterns

DeepMind London: Ivan Leo on Agents, and the People Building With Them

Jet Semrick on agent work, plus my Compile workshop workflow: capture notes in Notion, brief separate bots, and keep editorial judgement.

jet semrick: grok bot workshop | compile london jet semrick on agent work, plus my compile workshop workflow: capture notes in notion, brief separate bots, and keep editorial judgement. ai developer-experience tools

Jet Semrick: Grok Bot Workshop | Compile London

Tomas Reimers on code review after agents: validation, manageable changes, shared context, and reviewing the actual output.

tomas reimers: code review in the agent era | compile london tomas reimers on code review after agents: validation, manageable changes, shared context, and reviewing the actual output. ai software-development developer-experience

Tomas Reimers: Code Review in the Agent Era | Compile London

Pablo Santos Luaces session notes: two-way comparisons, three-way merges, best common ancestors, criss-cross history, and Origin performance.

pablo santos luaces: how file merges work | compile london pablo santos luaces session notes: two-way comparisons, three-way merges, best common ancestors, criss-cross history, and origin performance. software-development developer-experience tools ai

Pablo Santos Luaces: How File Merges Work | Compile London

Murali Suriar on diagrams as shared understanding: choose the right view, compare mental models, and review architectural change.

murali suriar: diagrams and shared understanding | compile london murali suriar on diagrams as shared understanding: choose the right view, compare mental models, and review architectural change. ai platform-engineering architecture-patterns observability

Murali Suriar: Diagrams and Shared Understanding | Compile London

David Gomes on agent worktrees: finish setup first, track where agents work, and make cleanup part of the harness.

david gomes: git worktrees for agents | compile london david gomes on agent worktrees: finish setup first, track where agents work, and make cleanup part of the harness. ai developer-experience software-development tools

David Gomes: Git Worktrees for Agents | Compile London

Pauline Brunet session notes: use-case selection, forward deployed engineers, production ownership, evaluations, permissions, and outcomes.

pauline brunet: ai deployments at scale | compile london pauline brunet session notes: use-case selection, forward deployed engineers, production ownership, evaluations, permissions, and outcomes. platform-engineering developer-experience software-development architecture-patterns ai

Pauline Brunet: AI Deployments at Scale | Compile London

Notes and photos from six Cursor Compile London sessions: AI deployments, worktrees, diagrams, merges, code review, and Grok Bot.

cursor compile london: six talks on working with agents notes and photos from six cursor compile london sessions: ai deployments, worktrees, diagrams, merges, code review, and grok bot. ai developer-experience platform-engineering software-development

Cursor Compile London: Six Talks on Working With Agents

On two Google repos, codebase-memory-mcp returned 22x smaller responses than disciplined grep, but its infra graph and MCP discovery fell short.

the token economy of codebase-memory-mcp: a knowledge graph benchmark on two google repos, codebase-memory-mcp returned 22x smaller responses than disciplined grep, but its infra graph and mcp discovery fell short. tools developer-experience ai

The Token Economy of codebase-memory-mcp: A Knowledge Graph Benchmark

AI agents have unlimited ambition and undefined access boundaries. You have sandboxed everything else. Here is how to sandbox your agents too.

ai tool gateways: sandboxing agent access in kubernetes ai agents have unlimited ambition and undefined access boundaries. you have sandboxed everything else. here is how to sandbox your agents too. security software-development kubernetes platform-engineering zero-trust architecture-patterns

AI Tool Gateways: Sandboxing Agent Access in Kubernetes

TTFT is your SLO, not throughput. Here is what to measure for LLM inference, and what Google ADK, LangChain, and LangGraph give you out of the box.

observing llm inference: the metrics that actually matter ttft is your slo, not throughput. here is what to measure for llm inference, and what google adk, langchain, and langgraph give you out of the box. observability software-development platform-engineering kubernetes architecture-patterns

Observing LLM Inference: The Metrics That Actually Matter

Network policies in most clusters are cargo cult: written, accepted, changing nothing. Cilium actually enforces them, and shows you the traffic.

network control with cilium and kyverno: policies that actually work network policies in most clusters are cargo cult: written, accepted, changing nothing. cilium actually enforces them, and shows you the traffic. kubernetes security platform-engineering multi-tenancy zero-trust architecture-patterns

Network Control with Cilium and Kyverno: Policies That Actually Work

Drop the service-account JSON key. Federate GitHub Actions into GCP with OIDC, then scope the service account to read one named secret, not the whole project.

keyless gcp secrets in github actions: workload identity, scoped to one secret drop the service-account json key. federate github actions into gcp with oidc, then scope the service account to read one named secret, not the whole project. gcp security ci-cd oidc zero-trust platform-engineering

Keyless GCP Secrets in GitHub Actions: Workload Identity, Scoped to One Secret

A Next.js static export on Cloudflare is fast, free, and has no server to patch: the whole setup, MDX to Terraform, in a clonable starter repo.

serving a next.js static site on cloudflare (like this one) a next.js static export on cloudflare is fast, free, and has no server to patch: the whole setup, mdx to terraform, in a clonable starter repo. cloudflare next.js platform-engineering developer-experience ci-cd

Serving a Next.js Static Site on Cloudflare (Like This One)

Your tenants want dashboards. Your security team wants isolation. Your SREs want a single pane of glass. Here is how to build all three with the LGTM stack.

multi-tenant observability: lgtm at platform scale your tenants want dashboards. your security team wants isolation. your sres want a single pane of glass. here is how to build all three with the lgtm stack. kubernetes observability platform-engineering multi-tenancy architecture-patterns

Multi-Tenant Observability: LGTM at Platform Scale

Kubernetes deprecated Ingress in spirit years ago. Gateway API replaces it with a model that actually separates platform from application concerns.

gateway api in practice: from ingress migration to envoy debugging kubernetes deprecated ingress in spirit years ago. gateway api replaces it with a model that actually separates platform from application concerns. kubernetes platform-engineering architecture-patterns developer-experience

Gateway API in Practice: From Ingress Migration to Envoy Debugging

kind runs real kubeadm in Docker, the same setup Kubernetes CI, Cilium docs, and Envoy Gateway quickstarts use. Why that matters for local platform work.

running local kubernetes with kind: ephemeral by design, production-honest by default kind runs real kubeadm in docker, the same setup kubernetes ci, cilium docs, and envoy gateway quickstarts use. why that matters for local platform work. kubernetes developer-experience platform-engineering setup tools

Running Local Kubernetes with kind: Ephemeral by Design, Production-Honest by Default

Why 48GB changes a Mac from coding laptop into an AI workstation for local models, parallel agents, and platform engineering.

the m5 pro setup: same mac, different era why 48gb changes a mac from coding laptop into an ai workstation for local models, parallel agents, and platform engineering. developer-experience macos setup tools software-development security

The M5 Pro Setup: Same Mac, Different Era

A practical GPG commit signing setup on macOS, with GitHub verification, pinentry-mac, GPG_TTY, and the parts that usually break.

git commit signing with gpg on macos a practical gpg commit signing setup on macos, with github verification, pinentry-mac, gpg_tty, and the parts that usually break. security developer-experience setup tools macos

Git Commit Signing with GPG on macOS

A clean SSH commit signing setup on macOS, with separate auth and signing keys, local verification, and none of the usual GPG hassle.

ssh git commit signing for busy engineers a clean ssh commit signing setup on macos, with separate auth and signing keys, local verification, and none of the usual gpg hassle. security developer-experience setup tools software-development

SSH Git Commit Signing for Busy Engineers

Authentication gets easier when you separate login, delegated access, SSO, and workload identity. This guide shows where each one fits.

authentication 101: a complete guide to modern identity methods authentication gets easier when you separate login, delegated access, sso, and workload identity. this guide shows where each one fits. security authentication oauth2 oidc saml sso kubernetes golang zero-trust mtls jwt

Authentication 101: A Complete Guide to Modern Identity Methods

Notes on Google's GAIL certification from a platform engineer: useful for strategy conversations, thin on the messy reality of shipping AI.

google's gail certification: useful strategy, thin engineering notes on google's gail certification from a platform engineer: useful for strategy conversations, thin on the messy reality of shipping ai. software-development gcp books platform-engineering developer-experience

Google's GAIL Certification: Useful Strategy, Thin Engineering

Run workloads on GKE and access AWS without static keys. This guide shows how to federate a Kubernetes service account into an AWS IAM role.

gke to aws identity federation: a guide to keyless access run workloads on gke and access aws without static keys. this guide shows how to federate a kubernetes service account into an aws iam role. kubernetes aws gcp security oidc platform-engineering

GKE to AWS Identity Federation: A Guide to Keyless Access

Why we added Terragrunt to Terraform to manage 100 plus services without copy paste, backend sprawl, or one giant state layout.

terraform and terragrunt: why we added terragrunt at scale why we added terragrunt to terraform to manage 100 plus services without copy paste, backend sprawl, or one giant state layout. terraform platform-engineering developer-experience

Terraform and Terragrunt: Why We Added Terragrunt at Scale

Seven books that shaped how I build platform products people trust, adopt quickly, and keep using.

7 books that shaped how i build platforms developers actually use seven books that shaped how i build platform products people trust, adopt quickly, and keep using. books software-development platform-engineering developer-experience architecture-patterns multi-tenancy

7 Books That Shaped How I Build Platforms Developers Actually Use

A deep dive into my zshrc configuration, featuring shortcuts for Kubernetes, Cloud Platforms, and AI tools to streamline your daily workflow.

boost your productivity: my zshrc configuration for platform engineers a deep dive into my zshrc configuration, featuring shortcuts for kubernetes, cloud platforms, and ai tools to streamline your daily workflow. developer-experience software-development tools platform-engineering

Boost Your Productivity: My Zshrc Configuration for Platform Engineers

Setting up a Mac for development - tools, configurations, and practices, updated as they change. "It's not much, but it's mine" - Every $5000 Setup.

the "it's not much, but it's mine" mac setup setting up a mac for development - tools, configurations, and practices, updated as they change. "it's not much, but it's mine" - every $5000 setup. macos setup tools

The "It's Not Much, But It's Mine" Mac Setup

Most Go client libraries are a pain to use, test, or extend. How to write one that isn't: functional options, context propagation, rich errors, retries.

writing a go client library worth using most go client libraries are a pain to use, test, or extend. how to write one that isn't: functional options, context propagation, rich errors, retries. golang software-development developer-experience architecture-patterns

Writing a Go Client Library Worth Using

A practical guide to deploying deep learning and computer vision models on Intel edge hardware with OpenVINO, from IR conversion to device trade-offs.

running deep learning models at the edge with intel openvino a practical guide to deploying deep learning and computer vision models on intel edge hardware with openvino, from ir conversion to device trade-offs. ai architecture-patterns software-development tools

Running Deep Learning Models at the Edge with Intel OpenVINO

Structured study notes from the SUSE Cloud Native Foundations scholarship. Covers cloud native design, Docker, Kubernetes basics, and open source PaaS.

suse cloud native foundations: my study notes structured study notes from the suse cloud native foundations scholarship. covers cloud native design, docker, kubernetes basics, and open source paas. kubernetes architecture-patterns ci-cd platform-engineering software-development

SUSE Cloud Native Foundations: My Study Notes

Cloud Run supports gRPC out of the box, once you understand how it handles TLS and HTTP/2. Here's the full picture, in Go.

grpc on cloud run cloud run supports grpc out of the box, once you understand how it handles tls and http/2. here's the full picture, in go. gcp golang developer-experience software-development