AI Systems Architect

Snapshot

The mission of 120ft Factory is to start the new western industrial revolution. Our core thesis is that the computational challenges of architecting a real-world autonomous factory are isomorphic to simulating massive, open-world factory games. We need a game engine for the real world.

Description

As the AI Systems Architect, you will view our factory not as metal and wires, but as a massive, real-time strategy game. Your mission is to build the 'Factory Tick', the core deterministic simulation loop that orchestrates the entire economy of our manufacturing cells. In simulation and game engine architecture, a «tick» is a single iteration of the update loop (e.g., happening 60 times a second). However, a Factory Tick is a specific, massively complex version of this concept designed to solve the «Many-Body» logistics problem.

You will apply Data-Oriented Design (DOD) and Entity Component Systems (ECS) to handle millions of discrete states, resource routing, queue prioritization, and agent pathfinding. You don't need to know how to bolt a robot together; you need to know how to orchestrate a stochastic, multi-agent world into a highly optimized, predictable logic engine.

• Architect the core deterministic simulation loop (the 'Factory Engine') from scratch, capable of handling complex dependencies, queues, and stochastic physical failures.
• Design the algorithmic 'economy' of the factory: modeling inputs, outputs, optimal agent pathfinding, and dynamic deadlock resolution in a constrained 3D space.
• Build the generative algorithmic bridge that converts raw human inputs (e.g., an product manual or CAD file) into an auto-generated, optimized dependency graph for the robots.
• Implement Data-Oriented Design (DOD) to decouple simulation time from rendering time, allowing us to run fast-forward predictions to identify bottlenecks before they happen in reality.
• Expand our logical simulation from a discrete 2D grid into a continuous, 3D spatial model that interfaces directly with our VLA AI models and physical hardware.

• Exceptional, production-grade programming skills in C++, with a track record of building highly complex, performance-critical logic systems or AI from scratch.
• Deep background in mathematics, competitive programming, or complex algorithmic problem-solving.
• Experience architecting game engines, complex AI for grand strategy games, or massive factory/logistics simulations (e.g., systems handling thousands of simultaneous agents/entities).
• Deep understanding of how to optimize CPU cache, state machines, and tick-rate stability.
• We do not care about formal degrees. Olympiad-level mathematics achievements or proven, elite-level competitive programming results are highly valued.
• You are a first-principles thinker. You are excited to take the math and logic used in gaming (like Factorio, Satisfactory, or Europa Universalis) and apply it to the reality of atoms.
• Experience with economy balancing, agent behavior trees, or high-density pathfinding (A*, RVO).
• You are comfortable with autonomy, taking a whiteboard concept and turning it into a hyper-optimized C++ architecture in a matter of weeks.
• High risk tolerance and an appetite for building something category-defining from the ground up.

We offer a competitive base salary paired with a equity grant for founding engineers.