Join Gen 7 Labs as our Founding Engineer to shape two working AI infrastructure products, own critical production decisions, and serve as the company's second qualified technical reviewer

What we are building

Gen 7 Labs is building infrastructure for reliable AI coding agents. Our first two products are working systems moving toward initial customer deployments, designed to deliver value independently or together.

Product # 1

Persistent project memory for coding agents. The Product captures durable decisions, conventions, constraints, and guardrails from agent interactions, resolves superseded guidance, and serves current project state across sessions through MCP or HTTP.

Product # 2

A governed platform for running coding agents against real repositories. The Product combines isolated worktrees, a kernel sandbox enabled in the production profile, human-gated write-back, and controls for cost, revocation, audit, and crash recovery.

The Product is built first for coding-agent memory. Coding agents do not reliably carry project context from one session to the next, so teams rely on hand-maintained repository instructions and documentation. Those artifacts grow, drift, and can keep superseded decisions in circulation.

The product turns agent interactions into structured, durable project knowledge: decisions, conventions, constraints, and guardrails. It resolves superseded guidance as that knowledge changes and serves the current project state through MCP or HTTP. Teams can use it with existing coding-agent runtimes or as the optional memory layer for the product.

Both products are already working and under active development. They have extensive automated testing and continuous integration. The opportunity is to shape the production architecture, harden the critical security and reliability boundaries, and help take the systems into their first customer deployments.

The role

- Architectural authority. You will be the second opinion on consequential design decisions and the primary decision-maker in many areas you own. Candid disagreement with the founder is expected.

- Full end-to-end ownership. You will take systems from design through deployment and operation: service code, containers, observability, incident response, and recovery. There is no separate platform or SRE team between you and production.

- Security-review authority. Several security decisions and threat-model sign-offs are intentionally waiting for a second qualified reviewer.



This hire will hold that responsibility and help turn reviewed controls into production evidence.

- Team building. You will help set the technical bar, shape the interview loop, and mentor the next one or two engineers. The role begins as a senior IC with no direct reports

What you will own

- Production infrastructure and operations. Build the GCP foundation end to end: Terraform, KMS, IAM, Workload Identity, deployment automation, backup and disaster recovery, and an on-call operating model.

- Security and reliability controls that fail closed. Turn tenant isolation, sandboxing, credentials, approvals, audit, and repository governance into enforceable production contracts. Replace checks that can pass vacuously with derived rules, negative controls, and live evidence.

- Trustworthy evaluation for probabilistic systems. Build corpus provenance, a sealed holdout process with human custody and a non-replenishable disclosure budget, human calibration, and statistical decision rules that can distinguish real improvement from noise - including the discipline to report "no result" when the evidence is insufficient.

- Scale and performance correctness. Redesign the product paths that grow with total tenant history and remove quadratic ingest behavior. Improve indexing, pagination, timeouts, and tenant-scoped query patterns across both systems.

- Product architecture across the full surface. Work on the coding-agent memory path - turn journaling, structured-claim extraction, supersession, and MCP/HTTP retrieval - alongside Postgres-backed orchestration, authentication and authorization, cost and audit, agent isolation, observability, and the operations UI.

Engineering systems and technical standards. Make branch protection, review, CI, documentation, release gates, and deployment posture enforceable. Help decide what should ship, what should remain disabled, and what should be removed.

Technical environment

- Product # 1- Python 3.12; typed synchronous services; PostgreSQL 16; pgvector/HNSW; Alembic; FastMCP; Docker; GCP Cloud Run.



The HTTP layer intentionally uses the Python standard library rather than a web framework.

- Product # 2 - TypeScript 6 / Node 24; NestJS on Fastify; PostgreSQL 16 with Drizzle; WorkOS; AES-256-GCM with GCP KMS; Next.js 16 / React 19; Docker cell topology; Claude Agent SDK; kernel sandbox enabled in the production profile, with proof hardening still in scope.

- INFRASTRUCTURE - GCP. Terraform, IAM, Workload Identity, deployment automation, backup/DR, and the production operating model are greenfield ownership areas. Kubernetes is not used today.

What we require

PostgreSQL at genuine depth is the central technical requirement. We mean schema and data-model design, reversible migrations against live data, query plans, locking and isolation levels, work queues, roles, and ideally hands-on row-level security. Both products rely on Postgres for tenant isolation; ORM familiarity alone is not enough.

- Production backend depth. You have shipped and operated real systems and can work deeply in Python and TypeScript/Node.js, or show solid evidence that you can cross that boundary quickly.

- Multi-tenant system experience. You have built or hardened authentication, authorization, tenant isolation, quotas, blast-radius controls, and data-boundary tests.

- Operational ownership. You have owned containers, deployment, observability, incident response, and recovery without assuming another team will finish the job.

- Product judgment in ambiguity. You have chosen what to build, not only how to implement a ticket. Seed/Series A experience, technical-founder experience, or another high-agency environment all count.

- Verification mindset. You think in failure modes, ask what happens when a process crashes halfway through, and prove controls by showing that removing them turns the evidence red.

- Clear written communication. Design records, change narratives, operational notes, and explicit limits are part of the engineering work.

- AI-assisted engineering fluency. We use coding agents heavily. You are comfortable reviewing, constraining, testing, and improving machine-generated code while remaining accountable for the result.

Equal opportunity and accommodations

Gen7 Labs evaluates candidates on role-related qualifications and welcomes applicants from a wide range of backgrounds. We provide reasonable accommodations throughout the application and interview process; candidates can tell the recruiter what would help them participate fully.

📌 Founding Engineer (Canada)
🏢 Gen7Labs
📍 Canada