Software Engineering LMTS / PMTS - Software Factory (REMOTE)

Engineering at Qualified is a high-ownership, high-velocity environment where engineers ship customer-facing product to real enterprises at scale. We’re building at the frontier of AI + product: developing agentic software and embedding AI directly into our platform, using modern toolchains and processes that help engineers move faster while staying focused on high-impact problem solving.
Following our acquisition by Salesforce, we continue to operate with the pace and ownership of a startup—backed by the distribution, resources, and long-term runway of a global enterprise. Engineers have true end-to-end ownership from idea to production: scoping, implementation, testing, deployment, monitoring, and iterative improvement based on real customer feedback, without layers of approvals slowing you down.
The Role
Software Factory is a new team that makes coding agents dependable in Qualified’s codebase and how work gets planned, implemented, shipped, and monitored. We run the best off-the-shelf agents (Claude, Codex, Cursor, and whatever comes next) rather than building our own.
This team's product is everything around those agents. That means a monorepo agents can orient in quickly, secure sandboxes where they run our full build, and a feedback loop that measures where agents stall or work from bad context, then fixes the cause.
As one of the team's first engineers, you'll shape that work from the ground up while shipping to demanding internal customers from day one: every engineer, product manager, and designer at Qualified.
What You'll Do
Make our monorepo agent-native: instruction files, skills, and progressive disclosure that get an agent to the right code in minutes, re-audited with every new model
Build the feedback loop: session telemetry across Cursor, Claude Code, and Codex, evals, and PR review analysis that separates wrong decisions from bad context
Own the agent sandbox: a full dev environment in a container, with proxy-only network access and secrets swapped at the proxy so agents never hold real credentials
Evaluate and integrate orchestration that starts agent work from Slack, Linear, GitHub, and schedules, with sound build-vs-buy judgment
Build the guardrails (verification, evals, and review) that let more agent work merge with less human hands-on time
Teach the engineering org to work with agents day to day, through enablement sessions and shared patterns
Own delivery from design through launch, monitoring, and iteration
Who You Are
A heavy daily user of coding agents (Claude Code, Codex, Cursor, or similar), with real opinions about where they shine, where they fail, and how to instruct them
Experienced building developer tooling or developer experience that other engineers depend on daily
Comfortable owning a containerized environment end to end, including its network rules and secrets handling
Able to turn messy signals (session logs, PR threads, eval results) into a specific fix and show it worked
Experienced leading engineers to deliver complex systems
Energized by ambiguity: a new team means the roadmap and conventions are still being written
Pragmatic about build versus buy: the ecosystem moves weekly, and you'd rather integrate something great than rebuild it
Why Join Qualified
Define how software gets built: This isn't one product's roadmap. It's how the whole engineering org builds everything else, and your leverage compounds with every team that adopts it.
Day-one greenfield: A brand-new team where the approach is yours to shape, with real internal customers and real constraints from the first week.
The frontier, applied: Coding agents are the fastest-moving area in software. This role is about making them effective, safe, and trustworthy at company scale.
High bar, low bureaucracy: Work with a team that values craft and decisive execution, where strong ideas win and shipping is the default.
If you're excited to make coding agents do real, dependable work for an entire engineering org, we'd love to hear from you.

Employment type
Full time