Founding AI Engineer [LLM Platform]

THE MISSION

Over $300B+ in complex travel bookings still depend on manual workflows. While the world moved to APIs, high-stakes bookings (weddings, conferences, group travel) are still handled by humans manually typing data from messy emails into legacy systems.

Lobby is the orchestration layer that fixes this. We are not building a chatbot; we are building AI agents that negotiate, quote, and execute transactions directly within the core infrastructure of global travel brands.

THE CHALLENGE

As our Founding AI Engineer, you do not "integrate an API and call it AI". You own the intelligence layer of a production system that enterprises depend on. You will:

• Run the Engine: Take ownership of our existing LLM orchestration engine and evolve it, from prompt architecture to state handling to failure recovery.
• Scale per Client: Set up and manage dedicated model instances and configurations for each enterprise client, and build the tooling that makes onboarding the next ten clients boring.
• Master Model Switching: Route between models and providers based on task, latency, and cost. Swap models under the hood without breaking client-specific behavior.
• Optimize Relentlessly: Own the cost and performance curve of every token we spend. Build evals that catch regressions before customers do.

YOUR STACK

We value boring technology used for exciting problems. We optimize for shipping speed:

• Core: Python (FastAPI/Django), PostgreSQL, Redis.
• The AI Layer: Multi-provider LLM integration, custom orchestration logic, prompt pipelines, eval harnesses.
• Infrastructure: AWS, Docker.
• Observability: Tracing, cost tracking, and eval dashboards you will help shape.

MUST HAVES

• Seasoned Builder: Minimum 8 years of experience as a software engineer (fullstack, or at least more backend / infra heavy).
• Startup DNA: You have worked at a startup before, in a small team, and you know what it means to ship without layers of process around you.
• Engineer First: You are a software engineer who transitioned into AI, not a data scientist who learned to deploy. You think in production systems, not notebooks. No PhD required or expected.
• Battle-Tested in AI: You have shipped and operated complex AI setups in production. Multi-step agent flows, per-tenant configurations, real users, real failure modes. Not just a thin wrapper around a chat endpoint.
• Evals as a Discipline: You treat prompt engineering and evals as engineering, with versioning, regression testing, and measurable outcomes, not vibes.
• Proven, Not Self-Declared: You can point to systems you have built and run, not just talks you have given. We hire evidence, not job titles.
• Swiss-Based/Local: You want to be on the ground in Zurich, building alongside the founders in a high-intensity environment.

WHY LOBBY?

• Extreme Autonomy: No "Middle Managers." No "Agile Coaches." Just builders, customers, and code.
• Market Pull: We are tackling a massive, un-sexy problem that is desperate for a modern solution. The "Why Now" is obvious once you see the product.
• Founding Equity: You get a seat at the table, not just a desk.

HOW TO APPLY

Do not send a formal cover letter. Send us:

• A link to an AI system you built or operated in production (repo, write-up, or demo).
• The gnarliest model, prompt, or scaling problem you have solved, and how you knew it was actually fixed.

Email: careers@lobbyai.io