Founding Engineer

Marble is the first AI-native operating system for the restaurant back of house, and the first agentic OS for physical industries, starting with multi-unit restaurant groups. Restaurants, cafes, and bars, a $1.5T industry, still run on disconnected point solutions, spreadsheets, clipboards, and tribal knowledge. That costs billions a year in stockouts, food waste, and lost profit. Marble replaces that stack. Computer vision and voice AI capture real-time inventory about 90% faster than manual counts. ML models forecast item-level demand at about 97% accuracy. AI agents run the workflows: procurement, invoice reconciliation, scheduling, food prep, menu engineering, and more.

The founders met at the University of Michigan. One is CS and engineering, formerly at Bloomberg, with five years building trading systems for hedge funds. The other comes from business school, and his family owns restaurants. It is YC-backed. In under four months it signed 150+ restaurants across the US, Canada, and Europe, and has grown 25% week over week for three months.

The technical bet is three stacked problems: capturing what is actually happening in the operation (computer vision, voice, documents, and integrations), forecasting item-level demand over noisy signals such as weather, game schedules, promos, and catering, and agents that make real-money decisions every day, what to order, how much to prep, who to schedule, and which invoice to dispute. The stack is GCP, distributed systems, NoSQL, event-driven architectures, REST APIs, model harnesses, and backend plus web and mobile. The team is in New York and ships to real kitchens within the week. Seed stage, about 1 to 10 people, founded in 2026.

Marble is hiring its founding engineer, effectively a second technical co-founder. With 150+ restaurants signed and deployment underway, the technical co-founder is the single technical point of failure. He needs someone who can take a vague task and drive it to done. You are a product builder: you drive the roadmap, talk with customers, sell the product, and have a voice in strategic decisions from day one. Some restaurants have enough custom flows that you may own a customer's workflows directly, so communication and comfort with customers matter.

The spike they want is product, not infrastructure. The technical co-founder covers infra. They want someone who thinks about users and driving value, with real taste, because Marble is not a technical product, it has to be tasteful. You build the full path from an order to the screen a manager approves. You own the agent harness: context, orchestration, persistent state, and human approval when an agent is moving real dollars. You take messy inputs (POS, accounting, payroll, invoices, recipes, vendor catalogs, stock counts) and turn them into clean ground truth. You make the agents dependable: evals from real operator workflows, failure tracing, recovery and escalation, and model choices based on production evidence. The early roadmap also includes paying vendors on Marble, and a proactive assistant for restaurant managers.

Bar: about 1 to 3 years, early-career, and AI-native. Strong fundamentals in Python or JavaScript/TypeScript, and willing to get productive in both. You have shipped agents that use tools and take actions in production, and you have built or debugged the harness: context, tool execution, state, failure recovery, and evals. You can discuss something substantial you built, including the trade-offs and what you would change. You notice messy details. LangGraph, LangChain, LangSmith, retrieval, SQL and NoSQL, and judgment about when an agent loop beats a deterministic workflow.

Founder mentality is the screen. Cool startup experience is strongly preferred, ideally early at something that scaled. A pure big-tech habit of waiting for a spec and leaving at five is what they filter against. A known school is a signal, not a Stanford-or-bust bar. Having founded something, or built something real, weighs more than a GitHub profile.

In office in New York, at an intense early-startup pace. Visa sponsorship is limited. Equity is about 1% to 3%. Health, dental, and medical. Posted work intensity is 9-9-6. The process is a short phone screen with the technical co-founder, a vague take-home that shows how you use AI tools to build something real fast, a review of that project plus an on-the-spot system design conversation with no AI, then a behavioral conversation with the co-founder.