Senior Engineer, Developer Productivity
Required
● 5+ years of writing and maintaining production code primarily Python. You
should be
comfortable in a 5K+ line Python codebase, idiomatic with typing, pathlib,
subprocess, pytest, and the subtleties of mocking at the import-site vs.
the source-site
(this will come up).
● Deep familiarity with Docker, docker-compose, and the Dev Containers
specification.
You should be able to read a devcontainer.json plus its referenced
docker-compose.yml and predict exactly which bind-mounts, environment
variables,
and lifecycle hooks will land inside the resulting container.
● Comfortable across macOS, Linux, and Windows as both a user and a developer.
Cross-platform shell quirks (PowerShell vs. bash, path-separator handling,
symlink
semantics, NTFS junctions vs. POSIX links) should not be a surprise.
● Strong Git: worktrees, hooks, submodules, the difference between `git
rev-parse
--git-common-dirand--git-dir, and exactly whatgit add -A` will
silently sweep into a
commit.
● A working understanding of at least one of: large C++ codebases
(CMake/Bazel,
distcc/ccache, NFS-backed build artifacts) or large Java codebases
(Maven/Gradle, IDE
indexing, build caches). You will not write application code, but you will
need to read it
and understand why a build is slow inside the container.
● A customer-support mindset. Engineers will ping you in Slack channels with
vague
reproductions while they are mid-sprint. You need to be the person who
replies calmly,
asks the two right diagnostic questions, and ships the fix the same day.
Nice to have
● VS Code extension API experience. UDE ships a small companion extension
under
vscode-extension/ude-helper/.
● Experience with Vault, NFS automounts, or corporate certificate
infrastructure.
● Experience maintaining a developer-facing CLI used by 100+ engineers.
● Familiarity with JFrog Artifactory (anonymous and authenticated access
patterns).
● Comfort owning a project end-to-end without a PM or designer.
Working style and what success looks like
This is a highly autonomous role. You would be the sole owner of UDE
day-to-day. Decisions
get made by you in conversation with users, not by committee. The first 30
days are about
learning the codebase, the dev-VM provisioning flow, and the user base; by 90
days you should
be shipping fixes faster than tickets come in.
You would report into the Production Engineering organization but spend most
of your time in
conversation with Purity, FA-UI, FlashBlade, and infrastructure engineers.
Expect Slack-heavy,
interrupt-driven days punctuated by 2–3 hour focused blocks for the harder
bug hunts and the
next-release-cycle features.
What success looks like at 6 months:
● The Jira backlog is leaner: same volume in, faster turnaround out.
● New engineers onboard onto UDE without help by reading the docs.
Stack
Python 3.9+, pytest, unittest.mock; Docker and docker-compose; VS Code + the
Dev
Containers extension; bash, zsh, PowerShell; TypeScript; macOS, Linux,
Windows; Vault, NFS, SSH/cert
authentication; Git (heavy worktree use); JFrog Artifactory; Jira Cloud and
Confluence; Slack;
GitHub Actions; MkDocs.
