Senior AI Integrations Engineer

The opportunity

The team behind Unity’s visionOS support, Play-to-Device, and Record & Playback is growing as we prepare for something even more ambitious: adapting PolySpatial to stream Unity content into other game engines and 3D environments — in-process, cross-process, and over the network.

We’re building systems that allow real-time 3D experiences to run, render, and interact in places they weren’t originally authored for, including other major game engines and emerging software platforms. We’re also developing new ways to encode and exchange interactive 3D content: protocols designed not only for rendering and interchange, but also for understanding, transformation, generation, and orchestration by future AI-driven tools and systems.

You’ll help invent the protocols, media formats, and runtime systems that bring interactive 3D content to novel platforms. Working at the intersection of core engine technology, graphics, networking, distributed systems, media, and AI, you’ll help define the interchange layer for the next generation of interconnected games and intelligent 3D ecosystems.

What you'll be doing

• Help define runtime protocols, formats, and execution models to encode, compress, and replay interactive 3D content between engines and AI models.
• Own and evolve critical subsytems across the engine stack, from low-level C++ runtime code to high-level C# APIs across editor, build, simulation, streaming, and playback.
• Build systems to capture, serialize, stream, and synchronize scene state, components, assets, user interactions, and rendering data under tight bandwidth, latency, determinism, and memory constraints.
• Design interfaces and feedback loops that allow AI systems to inspect engine state, invoke engine capabilities, and validate outcomes.

What we're looking for

• Core Engine Experience: Experience building or fundamentally extending game engine cores, middleware, or foundational systems (e.g. serialization, build pipelines, memory subsystems).
• AI Systems Fluency: Practical experience integrating AI/ML into production software, including areas such as model training, data collection, evaluation loops, agents, tool use, or model-context protocols.
• Architecture & Execution: Ability to design modular, testable systems - then implement, profile, debug, and ship them end-to-end.
• High-Performance Coding: Strong C++ and C# skills, with practical understanding of CPU caches, memory layout, allocation patterns, branch behavior, profiling, and performance tradeoffs in real-time systems.
• Engine-Wide Breadth: Hands-on experience across multiple engine domains — such as physics, animation, input, rendering, assets, or build systems — and how they interact.

You might also have

• Experience building infrastructure for training data generation, synthetic data, simulation environments, evaluation harnesses, reinforcement learning, or agentic workflows.
• Familiarity with 3D data formats and spatial content pipelines, such as USD, glTF, procedural generation, or semantic scene representations.
• Experience with MCP-style tool interfaces, LLM agents, multimodal models, inference pipelines, embeddings, vector search, or AI systems that inspect and act on structured application state
• Knowledge of compression, streaming, replication, authority models, RPCs, state diffing, delta encoding, or network protocols for real-time interactive systems.
• Experience with determinism, replay systems, simulation consistency, multithreading, job systems, lock-free programming, Unity DOTS, or other data-oriented runtime architectures.