Founding Engineer - AI Infra (Hybrid, Lviv)

Who you are

• 5+ years of backend / platform engineering experience
• You've shipped something end to end. A side project, indie product, or work project where you personally took it from idea to production and real users: architecture, code, deployment, and operations
• Daily user of agentic coding tools (Claude Code, Codex, Cursor, or similar), you build with them, understand how they call models and tools under the hood, and can reason about routing, configuration, and extending them
• Comfortable building and owning everything alone, you don't need a team around you to be effective. You can take a problem from design through implementation, deployment, and operations solo, and stay productive without structure, standups, or someone else unblocking you
• Comfortable with deep product ambiguity: there is no incumbent to copy in this category. We'll be designing the product and the architecture from first principles, choosing the stack, making design calls no one has made before, shipping fast, and owning outcomes end to end

Technical requirements

• Experience building or operating API gateways, proxies, or high-throughput request-routing services, hands-on with Nginx (or comparable reverse proxies), SSL/TLS (certificate management, termination, mutual TLS), and production traffic routing
• Strong in at least one of: Python, TypeScript, or Rust, and comfortable across 3–5 programming languages overall
• Production experience with a web framework (Django, Next.js, React, or similar)
• Solid database experience, production work with a relational database (PostgreSQL preferred, or MySQL/similar): schema design, query performance, and migrations on live systems
• Experience with vector databases (pgvector, Pinecone, Qdrant, Weaviate, or similar), embeddings, similarity search, and retrieval pipelines
• Strong cloud computing experience (AWS, GCP, or Azure): networking, load balancing, container orchestration, and running production workloads - you've operated real systems, not just deployed demos
• Hands-on experience with LLM APIs (Anthropic, OpenAI) and MCP (Model Context Protocol)

• Startup experience, you've worked at an early-stage company (or founded one) and know what shipping with a small team under real constraints feels like
• Experience building multi-tenant SaaS systems, tenant isolation as an architectural property
• Experience with auth and identity: OAuth2 / OIDC / SAML, especially Microsoft Entra ID (Azure AD) integration
• Experience with billing/metering systems (usage-based pricing, Stripe, etc.)
• Background in security products (DLP, EDR, SASE) or IT management tooling for service providers
• Experience with device management / endpoint agents (Intune, RMM tooling)

• Founding-level equity - meaningful ownership, not a token grant
• Competitive salary (discussed openly at the first call, benchmarked for senior/staff-level engineers)
• Direct work with the founder – no layers, no legacy code, you shape the architecture from day one
• A concrete, revenue-backed problem: paying customers and validated demand, not a research project
• Hardware/co-working budget and paid time off

For the first few months concentrated in-person time in Lviv is required while we get V1 built. Apartment and travel covered, and we're flexible on the shape.

We're a Y Combinator-backed startup building the AI infrastructure layer for the IT services industry: the platform IT providers use to deliver governed AI to the small businesses they manage, and to build and run AI agents for them on the same rails.

It's a genuine blue-ocean space, no incumbent has built this yet - while demand is already strong and validated, with small businesses paying thousands of dollars per month for AI delivered this way today.

The founder previously built an AI coding platform to 2M+ users. You'd join as the founding engineer at the exact moment the category is forming.

What you'll build

You'll be the first engineering hire, working directly with the founder and owning large parts of the platform architecture. The system has two tightly connected halves:

• AI access & governance infrastructure (the gateway)
• A multi-tenant AI gateway: an Anthropic/OpenAI-compatible API layer that popular AI tools (Claude Code, Codex, Gemini CLI, and others) are pinned to via managed endpoint configuration
• Identity binding: every request tied to a real employee identity via Microsoft Entra ID (OIDC/SAML federation), with per-tenant isolation across hundreds of end-client organizations
• Usage metering, pooled credit management, and per-tenant billing
• Audit logging, policy enforcement, and sensitive-data controls at the request layer
• An endpoint agent that deploys and pins tool configurations on managed devices and detects drift
• Agent runtime & builder infrastructure
• A tenant-isolated runtime where AI agents execute with scoped, vaulted credentials and inherit the permissions of the employee or role they act for
• A managed MCP (Model Context Protocol) server registry: sanctioned connectors to business systems, scoped per tenant
• Agent observability: replayable execution logs, approval workflows for actions, blast-radius controls
• The builder layer on top: templates and an authoring flow that lets non-experts assemble agents that are safe by construction

This is real infrastructure work: multi-tenancy, identity, credential security, gateways, and billing, with an AI-native product on top.