Software Engineer, AI Developer Productivity

Mirage•New York, NY

23h•$175,000 - $300,000•Onsite

About The Position

Mirage is hiring a backend engineer to rebuild how the company ships software as AI agents increase the volume of code produced. The role owns the path from commit to production, improving build, test, review, deployment, and security systems so engineers and agents can ship quickly and safely. The engineer will treat development infrastructure as a product: work directly with engineers to understand friction, measure bottlenecks, and build tools they want to use. The role also involves shipping features across Mirage’s AI video products and developer platform to gain firsthand experience with the systems being improved. Mirage says it has no hierarchical titles or levels internally and is currently hiring engineers who have operated at senior or staff+ scope. Most roles require in-person work at the NYC headquarters in Union Square.

Requirements

At least 5 years of professional experience building backend or full-stack software, with a record of shipping high-impact products or systems end-to-end.
Strong product sense and experience shipping products users value; apply the same attention to engineers and agents as users of development systems.
Hands-on experience with CI/CD, cloud infrastructure, or security, and motivation to develop deep expertise; specialists and product engineers ready to specialize are welcome.
Extensive use of coding agents in current work and informed opinions about their limitations.
Strong security instincts, including considering how systems could be attacked during development.
Practical, measurement-driven approach: measure before fixing and favor effective solutions delivered promptly over waiting for a perfect solution.
Comfort owning build, deployment, review, and security together, working across layers, and defining a roadmap where no playbook exists.
Interest in staying close to product, shipping to production weekly, and using firsthand experience to guide infrastructure improvements.
Preference for making the right path easy to follow through useful tools and thoughtful defaults.

Nice To Haves

Product engineering experience with recurring involvement in build, deployment, or security problems and an interest in focusing on those areas.
Experience building internal developer platforms or productivity tooling at a fast-growing startup.
Offensive security experience, such as penetration testing, red-teaming, bug bounties, or CTFs.
Experience shipping LLM agents or evaluations in production.
Experience with merge queues, monorepo tooling, or test-impact analysis.
Experience as a startup founder or early engineer.

Responsibilities

Ship features with product engineers across video creation software, the video rendering engine, and the developer platform, identifying development-system bottlenecks.
Build tools, agents, and streamlined workflows that help engineers ship quickly.
Rebuild CI/CD to support substantially higher change volume through incremental builds, test selection, flaky-test detection and quarantine, merge queues, and preview environments.
Improve deployment safety with progressive rollouts, automated canary analysis, feature flags, and automatic rollback.
Own environments for coding agents, including sandboxes, ephemeral stacks, and tools agents need to build, test, and verify work.
Build an AI review pipeline with codebase-tuned agent reviewers, risk-based routing to human reviewers, and automated merge-gating policy checks.
Run continuous agent-driven penetration testing against applications, APIs, and infrastructure, and turn findings into proposed-fix pull requests.
Integrate secrets, dependency, and supply-chain scanning into development workflows; provide AI-assisted vulnerability triage and automatically generated fix pull requests.
Secure agents’ access to code, infrastructure, and data using least-privilege permissions, short-lived credentials, sandboxing, audit logs, and prompt-injection guardrails.
Measure and improve the development lifecycle, including reducing PR-to-production time, replacing long-lived credentials in CI and agent environments, and reducing flaky tests through detection, quarantine, and fix pull requests.