Spatial & Game Designer developing Gen AI Mythologies and Human-AI Collaborative Pipelines / 25 years building environments, now designing how people direct AI to build them
I am Don Poole, artist, designer and veteran AAA/VR/AR developer. I currently design methodologies and tools that let people direct generative AI with precision, predictability and consistency: drawing out what a model knows about a place, then confirming, correcting and building on it. The work builds on 25 years of building simulated real-time environments at scale, with significant contributions as a world designer and builder to award-winning products from Naughty Dog, Sucker Punch, Wevr and Another Axiom that millions of people experience worldwide.
World Labs Marble with single Gemini 3.1 Flash Image generated from prototype multi-stack domain system instruction. Pure domain search, retrieval and synthesis. No image/multimodal prompting.
My foundation is unusual for this space. I studied photography and mixed media at the University of New Mexico under renowned photographers Patrick Nagatani, Tom Barrow, and art historian Eugenia Parry Janis, thinkers who understood that artists don't follow the evolution of their medium, they drive it. That conviction has shaped how I approach every technological shift in my career, from real-time 3D to VR and AR to spatial AI. I see it as the next phase of a conversation about how we make, see, and experience the world, one that, for me, started with photography and that continues to evolve as I do.
Images generated from text alone: what an image model knows about one neighborhood in New Town, Edinburgh.
LSA extraction and synthesis using older manual pipeline. LSA System Instruction (SI3.9) in Gemini 3.8 Flash extracts and syntheses deep domain specifics to generate prompt for Gemini 3.1 Flash Image (Nano Banana 2). Sessions included additional in chat prompts to move vantage points testing spatial stability within a chat session. User prompt: location, minimal scene description brief and asked "What do you see". Follow up prompts ask to move to other locations and describe what is seen. The text description from the multi-domain stack synthesis writes the prompt, not the user No image/multi-modal or any other input. No Grounding with Google Search, Image or Maps.
When I started my career, real-time 3D was genuinely new, not a mature industry with established pipelines, but a frontier where every project required inventing the process alongside the product. That early experience of operating without a playbook became a through-line: VR before it had a mainstream audience, mixed reality before the hardware was reliable, and now spatial AI before anyone has fully agreed on what it means. Twenty-five years later, the tools have changed completely. The instinct hasn't.
Ainslie Place Garden, Edinburgh, UK. Pure LSA extraction and synthesis using manual pipeline. LSA System Instruction (SI3.9) in Gemini 3.8 Flash extracts and syntheses deep domain specifics to generate prompt for Gemini 3.1 Flash Image (Nano Banana 2). 3 additional in chat prompts to change seasons. User established a location and scene brief. Subsequent prompts test in-session model stability with seasonal changes. No image/multi-modal or any other input. No Geo-Data grounding used. No grounding with Google Search, Image or Maps.
My current work is a method I call Latent Space Archaeology. It draws out knowledge in a model’s training that everyday prompting isn’t structured to find. The model is questioned through expert lenses (geology, civil engineering, architecture, botany, meteorology), each aimed at one real place, and the answers are combined into a detailed account of that place: its rock and soil, its plants, its weather, its history. That account goes in front of a director to check and approve.
LSA Studio and its render pipeline are the tools I’m building to apply the method: the model proposes, the director confirms, and the director’s word wins. I design the method and the tools, set the requirements and test; Claude writes the code from my designs.
"Here's your location. What do you see?" 1st gen (06-2026) "Eye" system instruction for walking tours w/image output. Images generated using domain stack/latent space engine in Gemini Flash 3.5.. Manual system instruction prototyping. The prompt is then dispatched to Gemini Flash 3.1 Image (Nano Banana 2). Images shown here are from a test of sequential prompts in the same chat session to maintain the spatial grounding as the view moves along the Water of Leith from St. Bernard's Well to Dean Bridge, Edinburgh, UK. User inputs a world coordinate only and asks "What do you see?". No multi-modal prompting (no actual photographs, photogrammetry of AI generated images used at any point in the pipeline.
From AAA to Emerging Technology / 25 years of evolution
Another Axiom
Another Axiom: Gorilla Tag - Pod Captain & Game/Level Designer
Mar 2023 – Nov 2025
Gorilla Tag has 15 million players and one of the most genuinely organic communities in all of VR. People don't just play it, they build culture around it. Working on it meant taking that seriously at every design decision.
I owned the full product lifecycle for 5+ new levels and led development of new locomotion pods (Swimming, Gliding, Hoverboard), each requiring its own spatial layouts and player physics framework. I also overhauled the game hub and store layout, improving navigation and item discoverability, and established visual and gameplay metrics for VR spatial design that the team could reuse across future work.
Wevr
Wevr / Jon Favreau: Gnomes & Goblins: Sr. Designer, Level & AI Interaction followed by Lead Designer prototyping VR experiences - Aug 2020 – May 2022
Nominated, Best VR Experience, Venice Biennale Film Festival 2020. Designed level layouts and AI interaction systems for goblin-player encounters, using Unreal 4, Blueprints to orchestrate WEVRs proprietary AI component system. Created 20+ unique AI events to simulate a living, reactive society.
Subsequently led design for a cross-platform Metaverse prototype and created a 3 point (HMD, L/R controller) player pose and interaction inputs (Unreal 4, Blueprints) for an unannounced Quest 2 title
Independent R&D: Unity, Project Tango, AR Core and ML Agents
Independent AR and ML R&D - Jul 2017 – Nov 2019
Designed a mixed-reality character framework in Unity using ARCore for real-time spatial awareness and procedural non-verbal communication.
Created real-time shadow approximation, world object occlusion and nav-mesh navigation using Unity and ARCore.
Scripted character interaction and response systems driven by user action and proximity, with Mecanim animations and blendshapes.
Conducted hands-on experimentation with Unity ML Agents, developing a practical understanding of reinforcement learning in spatial environments.
Rascali and Studio Roqovan
Rascali / Studio Roqovan | Lead Designer VR/AR (Early employee, equity stakeholder)
Jan 2015 – Jul 2017
Studio Roqovan (formerly Reload Studio) — joined as an early, equity-holding hire during the studio's founding period. Defined the creative product vision, UX strategy, and development roadmap for B2B educational VR applications. Improved customer ratings from 1.5 to 3.75 stars on a key Gear VR product through targeted design and interaction overhaul. Directed production reviews with clients including The American Experience, The Rubin Museum, and Boulevard VR.
Built multiplayer levels for World War Toons (PS VR) and pioneered diegetic 3D menu systems to resolve VR-specific UI challenges.
AAA World Design: Naughty Dog, Sucker Punch, Studio Roqovan/Rascali, Wevr, Another Axiom
AAA World Designer
I was a Level Lead at Naughty Dog for 7 years where I built and maintained assigned levels working as the contact point for art, animation, code and QA.
At Sucker Punch I was the Senior Environment Designer. I designed and blocked-out massive open-world game spaces that showcased new characters, enemies and gameplay features. I led the development of a fully destructible environments that allow the player to create the excitement and not a cutscene.
I was one of the 1st hires at Kojima LA and was part of the core group that recruited and built the L.A. team through it's founding 1st 18 months. I led multiplayer open-world map design and led the post-launch content and roadmaps for DLC to sustain player engagement.