Skip to content
View 4thel00z's full-sized avatar
🧂
don't be salty
🧂
don't be salty

Block or report 4thel00z

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
4thel00z/README.md

Product Engineer (AI)

I build and ship product features end-to-end: from idea → architecture → implementation → shipping → iterating with real-world customer feedback.

What I build

  • AI-assisted product features (agents, retrieval, evaluation workflows)
  • Backend APIs and services (clean contracts, pragmatic architecture)
  • Event-driven systems and integrations
  • Observability-first systems (metrics, logs, traces; debuggable by default)
  • Developer experience (SDKs, tooling, paved roads)

How I work

  • Strong ownership, fast iteration, and clear communication
  • prefer boring, reliable solutions (and I know when not to)
  • care about maintainability: tests where they matter, docs where they help
  • like tight feedback loops: instrument, ship, observe, improve

Current interests

  • Product engineering in the AI space (agents, retrieval, evals, guardrails)
  • Building systems that are simple to operate and hard to break

Toolbox

  • Core: Python, Go, Rust, TypeScript
  • Data: Postgres, MySQL, SQLite, Redis, Qdrant
  • Infra: Linux, Docker, Kubernetes, Terraform, Helm
  • Messaging & async: Kafka, RabbitMQ, SQS/SNS, Temporal, DBOS
  • Cloud: AWS (primary), plus Azure/GCP when needed
  • Observability: OpenTelemetry, Prometheus, Grafana, tracing/telemetry-first debugging
  • AI systems: RAG/search workflows, evaluation harnesses, retrieval quality loops, guardrails

📬 Contact

Open to product engineering roles at early-stage startups (especially in the AI space).

Recent projects

  • slurp — Cross-Document RAG eval dataset generator with Confluence crawler and LLM-powered question generation using clean architecture • Python — latest release: v0.1.0 (2025-10-25)
  • metalbrew — 🍺 Homebrew rebuilt in Zig — a dependency-free bottle package manager for macOS arm64. No Ruby runtime, hexagonal architecture, single static binary. • Zig
  • modal-skills — Claude Code skill teaching the Modal CLI — serverless Python & GPUs in the cloud (run/serve/deploy, volumes, secrets, shell).
  • s5cmd-skill — Claude Code skill for s5cmd — blazing-fast, massively parallel S3 data movement. Teaches the agent every s5cmd subcommand, grounded in v2.3.0 --help.
  • 4thel00z — My github frontpage.
  • memories • Go
  • konan — Blazingly fast text chunkers in Rust with pythonic bindings. Rayon-parallel, real async, char-accurate offsets, semantic chunking • Rust — latest release: v0.2.4 (2026-06-07)
  • ytdown • Rust — latest release: ytdown-cli-v0.3.1 (2026-06-06)

Pinned Loading

  1. slurp slurp Public

    Cross-Document RAG eval dataset generator with Confluence crawler and LLM-powered question generation using clean architecture

    Python 2

  2. lambda lambda Public

    🧰 Functional programming library for Go — ✅ Option-based error handling, 🔗 fluent pipelines, ⚡ parallel + channel utilities (with generics)

    Go 4

  3. webhookd webhookd Public

    Why trust online webhook services with your data when you can run your own webhooks? 🎣

    Go 5

  4. percently percently Public

    ⚡️ Blazing-fast percentile calculator (p50/p95/p99) in Rust — with Python bindings

    Rust

  5. konan konan Public

    Blazingly fast text chunkers in Rust with pythonic bindings. Rayon-parallel, real async, char-accurate offsets, semantic chunking

    Rust

  6. write-well write-well Public

    A Claude Code skill for writing prose that doesn't sound like AI wrote it. A tight reference an agent can actually apply: four cuts, a five-step edit pass, red flags, and a full 100-principle canon…