5-7 years of experience in software development, particularly with TypeScript and related technologies.
Demonstrated ability to ship products or features, evidenced by a portfolio or open-source contributions.
Experience in writing evaluations and comparing models in a technical context.
Proven track record of automating workflows and ensuring their reliability and observability.
Strong opinions on technical design and product direction, with the ability to defend them.
Responsibilities
Ship features from design to production that enhance customer experience.
Engage with customers to gather insights that inform prioritization and decision-making.
Develop agents that utilize tools effectively, ensuring transparency in their processes.
Write evaluations and validate outputs, navigating complex scenarios.
Automate manual workflows, incorporating verification and traceability.
Enhance existing systems, including model harnesses and coding agents.
Benefits
In-office work four days a week with one flexible remote day.
Equity options in a rapidly growing startup.
Comprehensive health, dental, and vision insurance.
401K plan with company matching.
Generous paid time off and all federal holidays observed.
Supportive team culture with collaborative values.
Dinner provided for late work hours.
Full Job Description
What you'll do:
Ship features customers use every day, from design through production.
Talk to customers and use their insight to guide your prioritization and decision making.
Build agents that call tools, cite their sources, and show their work.
Write evals and verify outputs that don't always seem black and white.
Take workflows people run by hand and make them autonomous, with verification, traceability, and observability built in.
Improve how we build: our model harness, our coding agents, our infra, etc.
Who you are:
Everyone we hire:
Has high agency. You see a problem and fix it without being asked.
Has a bias to action and works with urgency.
Brings real effort and drive; we can teach you Typescript, we can't teach you to care.
Cares about America and her interests.
Things we look for:
You've shipped things we can look at: open source, side projects, a portfolio, or something you launched publicly at a company.
You've written an eval, compared models against each other, or built with coding agents.
You've taken a loop a person was running and made it mostly run itself. You can show how you know it works.
You might be a deep specialist in evals, infra, testing, or systems.
Going above and beyond:
You've owned a large part of a product, or made high-impact contributions across several product areas.
You've been a lead, done well at a startup like ours, worked on a stack like ours, or written code at big-tech scale.
You have strong opinions on technical design and product direction, and you can defend them.
You're fully self-directed. You may lead other engineers.
Our stack:
Our primary app is written end-to-end in TypeScript and Effect, React and Next.js on the front end, and some Rust in our infrastructure (e.g. agent sandboxes).
Prisma, tRPC, Postgres, Redis, and Temporal for background workflows.
Models from OpenAI, Anthropic, and Google, run through our own harness.
Cursor, Linear, Slack, and Google Workspace.
Access to Claude Code, Codex, Devin, Grok Bot, etc.
Where you might work:
Capture: finding and qualifying opportunities, and managing pipeline.
Proposal: drafting and reviewing content with AI that cites its sources and satisfies compliance requirements.
Contract: deliverables, modifications, and past performance management for active contracts.
Platform: Dash itself, its sub-agents, automations, and integrations like Microsoft Graph, Salesforce, and Google Drive.
Data Library: how customer documents and records feed every AI workflow, with the user in control.
Logistics and comp:
In office 4 days a week in Flatiron with 1 flex day from home.
