What I use

A snapshot of the tools that show up most often when I’m shipping infra, reviewing code, or building agent-oriented tooling. This isn’t sponsorship — just what works for me day to day.

Skills & stack

Grouped tags are aligned with the public resume PDF (languages, platforms, and named tools across roles and open-source work). The sections below are narrative context, not a second list.

Languages

  • TypeScript
  • Rust
  • Go
  • Solidity

Kubernetes & reliability

  • Kubernetes
  • Helm
  • Istio
  • Argo CD
  • Flux
  • Karpenter
  • Chaos Mesh
  • K6
  • PagerDuty

Cloud & data

  • AWS
  • GCP
  • IBM Cloud
  • Cloudflare
  • CockroachDB
  • DragonflyDB
  • ClickHouse
  • Kafka
  • SQLite

Delivery & IaC

  • GitHub Actions
  • Jenkins
  • Terraform
  • Ansible
  • Packer
  • Pulumi
  • Docker
  • Docker Notary

Observability

  • Datadog
  • Prometheus
  • Grafana
  • Jaeger
  • Honeycomb
  • CloudWatch
  • LogDNA
  • Sysdig

Security & identity

  • Vault
  • Twistlock
  • Wiz
  • Okta
  • Rapid7

AI, agents & knowledge

  • OpenAI
  • Onyx
  • Apache Flink
  • Apache Tika
  • Gotenberg
  • Stirling-PDF
  • Paperless NGX
  • OpenRouter
  • Tavily

Blockchain & Web3

  • The Graph
  • Chainlink
  • Ethereum
  • Polygon
  • Slither
  • Manticore
  • Echidna
  • Stellar

Protocols & integration

  • GraphQL
  • gRPC
  • Protobuf
  • MCP
  • REST
  • Streamable HTTP

Development

  • Editor and terminal hygiene

    A modern IDE with inline AI assistance for fast iteration across TypeScript, Rust, and Go; paired with ripgrep, git worktrees, and scripted workflows for repeatable builds.

  • Kubernetes & GitOps

    Daily familiarity with EKS-style clusters, Istio-style routing, Argo CD / progressive delivery, and treating manifests and policy as code.

  • Observability stack

    Datadog, Prometheus/Grafana-style metrics, tracing (Jaeger/Honeycomb patterns), and structured logging for debugging distributed systems.

AI & agents

  • MCP and API ergonomics

    Building and dogfooding MCP servers (like ClawQL) where the goal is lean context: discover operations from bundled graphs instead of pasting giant OpenAPI blobs into prompts, with durable vault memory and optional add-ons (sandbox runs, schedules, Slack) when agent loops need more than read-only API calls.

  • Local experimentation

    On-device ML / GenAI samples and forks (e.g. gallery-style apps) to understand inference constraints and UX outside of pure cloud APIs.

Hardware

  • MacBook Pro + external display

    Primary driver for builds, container workflows, and long incident sessions — nothing exotic; reliability of the toolchain matters more than the logo on the lid.