Founding Engineer — TrustedRouter
HIRING
Founding Engineer — TrustedRouter
I need an engineer who can run the full stack of TrustedRouter with me. We sit between your app and 24 LLM providers — Anthropic, OpenAI, Gemini, Mistral, Cerebras, DeepSeek, Kimi, Together, Grok, Novita, Phala, SiliconFlow, Tinfoil, Venice, and more — across 7 regions and 2 clouds, with BYOK + KMS envelope encryption end-to-end. We're the LLM API your CISO doesn't have to worry about, and we're 100% radically open source (Apache 2) end to end.
You'd own the engineering operation end-to-end, working directly with me. That means the router itself — Python on Cloud Run across 10 regions, Spanner ledger, Bigtable logs, internal-gateway authorize/settle pattern across regions. It means the provider catalog: adding and removing providers, model aliases, keeping pricing fresh (we self-heal hourly via a GHA workflow that runs an LLM over public price pages). It means the trust story — BYOK key envelope, attestation-gated inference, the trust.trustedrouter.com page that customers' security teams actually read. BYOK + Stripe billing, OAuth, the console UX, the marketing site, prod smoke from GHA, Sentry, structured-log analytics, the synthetic monitor that runs in every region every minute. And the roadmap — you and I set it together.
The stack is Python 3.12 / FastAPI / httpx / Pydantic with strict typing and no ORMs. Cloud Run + Cloud Run Jobs, Spanner, Bigtable, KMS, Secret Manager, Cloud Scheduler — all multi-region. Deploys are gcloud CLI + bash: read the script, run the script. No Terraform, no Helm. GitHub Actions for CI, smoke, and self-heal. Sentry for errors, Axiom for structured-log analytics. Cloudflare in front, Stripe for billing, BYOK with envelope encryption per workspace. If you read that list and thought "good, I know exactly what each of those does and why" — keep reading.
What I'm looking for: at least 4 years shipping production systems you actually operated. You've owned an on-call rotation. You've run a service that customers paid for. You use LLMs every day in your own work — you know which provider is faster for what, which one's API is the worst, and what tool_use actually looks like in JSON. You can read a 700-line storage_gcp.py, find the bug, and ship the fix the same afternoon. You're comfortable being technically wrong in front of me and changing your mind. Bonus if you have experience with multi-region distributed state, multi-cloud, KMS / attestation / confidential compute, payments infra, or shipping a developer-API product to engineers — any one of those.
LLM routing is the part of the AI stack that's about to consolidate. Whoever builds the trust layer security teams approve — BYOK, audit, attestation, region pinning, no prompt retention — wins the enterprise tier. We have the regions, the BYOK, the provider catalog, and a CEO who's done the 0→exit motion before. I need the engineer who runs the rest of the way with me.
I'm Joseph Perla — exited CEO, early engineer at Facebook, Lyft, and Terminal.com.
Compensation: $100k–$300k salary, your pick along the cash/equity dial. 1–4% equity for the right hire — 4-year vest, 1-year cliff, double-trigger acceleration on acquisition. You'd be engineer #1. Remote, async-first, with whatever cadence of in-person works for both of us. Direct, fast feedback loops — no layers, no committees, no "design review forum."
To apply: sign up at trustedrouter.com, send a request to one of the providers, and email jobs@jperla.com with the model you called and what it returned, one thing about the product or stack you'd change in your first week, and a link to something you've built or shipped that you're proud of.
More on TrustedRouter
TrustedRouter — one API, all the LLMs, provably private.
Enjoyed this essay?
Follow me for more insights on technology, startups, and the future.