man okonkwo
about
i'm a senior systems engineer working on distributed storage and the runtimes underneath it. eleven years in, i still think the best code is the code that fails predictably. i've built ingest pipelines that survived a datacenter loss, replaced a homegrown queue with a raft cluster that hasn't paged anyone in two years, and mentored a dozen engineers into their first on-call rotations.
outside the terminal: bouldering, a slowly growing collection of mechanical keyboards, and a standing rule that i'll read any incident postmortem you send me.
cat experience.log
- 2022 — present
staff engineer, storage — meridian data
own the replicated log powering the multi-region object store; cut p99 write latency 4x.
- 2018 — 2022
senior engineer — northwind labs
rebuilt the event ingest tier in rust; migrated 40 services off a fragile queue.
- 2015 — 2018
backend engineer — cobalt systems
shipped the first version of the metrics pipeline that still runs today.
cat uses.md
[ editor ]
neovim + tmux, gruvbox material dark
[ terminal ]
wezterm, jetbrains mono, starship prompt
[ os ]
arch linux + hyprland on a framework 16
[ tools ]
ripgrep, fzf, atuin, lazygit, just
git log --oneline
- a1f9c2espoke at systemsconf 2026 on consensus under partition
- 7be04d1maintainer of the raft-rs merge queue since 2024
- 3cc812apublished 'the retry budget lie' — 40k reads
- 0d5a7f4first open-source commit to litewal