Software Engineer, Strategic Projects

About Keystone AI Keystone AI is building the new Internet by delivering software that makes it easy to securely interconnect people and their devices, no matter where they are. From hobbyists to multinational corporations, teams of every size use Keystone AI each day to protect their networks, share access to internal tools, and more. We're building a future for the Internet that's easy, sensible, and safe, like it used to be. Founded in 2019 and fully distributed, we're backed by Accel, CRV, Insight, Heavybit, and Uncork Capital. Job Description Join our Strategic Projects team, reporting to the team's Engineering Manager and ultimately into our co-founder and Chief Strategy Officer. This team exists to show developers and organizations why Keystone AI should be their default connectivity substrate, both for internal infrastructure and for embedding into customer-facing products. Aperture, our identity-aware LLM gateway, is one such initiative. This is an AI-native team. We use coding agents heavily as a force multiplier and expect every engineer to be a power user. Our customers demand a high bar on code quality, security, and operational rigour, and we treat that as a reason to use every tool available, including LLMs and coding agents, to ship fast without compromising it. We move fast from idea to prototype, and we're growing the side of the team that's energized by the path from prototype to dependable product. You'll work shoulder-to-shoulder with product-minded engineers from the moment an idea takes shape, shaping how it's built so it holds up later, and co-owning the work of making it production-grade: stable, observable, maintainable, and ready for real customer load. You're the kind of engineer who finds the long tail of development genuinely satisfying, the edge cases, the failure modes, the tests and runbooks, the details that turn a sharp demo into something customers can rely on. You'll work directly with reference customers on integrations and own the production bar on prototypes that graduate into Keystone AI 's product surface area, ensuring the security, review, and operational practices that wrap around agent-generated code actually hold in production. Keystone AI is fully remote, this team operates with significant ambiguity, and the shipping cadence is fast. Comfort working asynchronously and collaborating with highly technical customers is essential. Key Responsibilities Partner with product-minded engineers from the start, shaping prototypes (MCP gateways, LLM proxies, sandboxes, remote servers, and similar emerging workloads) so the path to production is built in, not bolted on. Co-own the journey from "works in a demo" to "stable in production": hardening, error handling, observability, and operational tooling. Own the production bar on prototypes that graduate into Keystone AI 's product surface area, including the security and review practices that wrap around agent-generated code. Work with reference customers to integrate Keystone AI , both for internal adoption (VPN replacement with identity-aware networking) and for embedding into their products, with particular attention to reliability, maintainability, and long-term supportability. Find and close the long-tail gaps, edge cases, regressions, flaky behavior, missing tests, undocumented operational knowledge, that separate a prototype from a dependable product. Establish and maintain the review, testing, and security practices that let the team ship agent-generated code fast without compromising the production bar. Light travel required ( What We Are Looking For Ideal candidates will have experience taking software from prototype to production in startup-like environments. We don't expect expertise with every aspect listed below. Let us know where you feel you have the most experience and where you'd like to keep developing. Strong at hardening, stabilizing, and maintaining production software; you take satisfaction in detail-oriented work and in closing the long tail of bugs, edge cases, and operational gaps. Go experience is a must; comfort integrating third-party APIs; basic networking. Heavy daily use of coding agents (Claude Code, Cursor, or similar) and LLMs in your engineering workflow, with a point of view on where they accelerate work, where they fail, and especially how to wrap them in the review, testing, and security practices that hold the production bar. Enjoy collaborating closely with exploratory, product-minded engineers and jointly owning the work of getting something solid into production. Comfort with the cadence of a strategy team at an early-ish startup: priorities shift with market signal, prototype-to-production timelines are short, and the bar on what ships stays high regardless. Thrive working independently in a remote environment. Nice to Have Experience operating production services: observability, on-call, incident response, performance and reliability work. Validating prototype code for production use alongside