Overview
This studio offers a hands-on approach to how humans communicate with computers and inhabit physical and virtual space. You will build works of self-expression and interaction across tangible, digital, and expanded realities — and ship them to headsets and to the LED wall.
Read the full Fall 2026 syllabus → — the whole term in one page: grade breakdown, the fifteen-week grid, every project spec, and the course AI policy.
What's new this year
Unreal Engine 5.8
The final UE5 release before UE6. Megalights and Dataflow are now production-ready, Lumen Lite targets 60 fps GI, and experimental Mesh Terrain, Toon Shader, and MetaHuman Collections open new territory for environments and crowds.
Agents in the editor
UE 5.8 ships an official in-editor MCP server, which means AI coding agents can spawn actors, wire Blueprints, build materials, and run tests inside your project. We will use Claude Code and OpenAI Codex as studio assistants all semester.
Virtual Production
The Cass Building VP Studio remains the capstone venue: LED cabinets, camera tracking, nDisplay. You'll learn to run the volume from the video server down to troubleshooting the wall.
Course objectives
- Understand the phases of Unreal Engine / Twinmotion level design from concept to completion.
- Become proficient in architectural modeling and level design, at true 1:1 scale with real environments.
- Build interactivity into UE levels with Blueprints — and direct AI agents to help you build it faster.
- Use Claude Code and Codex responsibly: prompt, review, verify, and document AI-assisted work.
- Package and deploy UE5 projects to Meta Quest and HTC Vive headsets via OpenXR.
- Operate the Virtual Production Studio from the video server to the LED cabinets.
- Develop your own voice in virtual and mixed-reality environments.
A Friday in studio
We meet once a week for four hours. Every session follows the same rhythm:
Demo + concepts. New engine territory introduced on the projector.
You build the same thing at your station, step by step.
Weeks 1–8: supervised practice on your own project, no new material. From Week 9 the same block becomes the agent lab — same problem, agent-assisted: prompt Claude Code or Codex, review what it did, fix what it got wrong.
Screens up. Show progress, trade techniques, plan the week's homework.
Grade breakdown
| Component | Weight | Notes |
|---|---|---|
| P1 · Digital Twin Flyover | 10% | Weeks 1–3 · Cesium, at true coordinates · due W3 crit |
| P2 · Architectural Visualization | 15% | Weeks 3–6 · built at 1:1 and sited on the globe |
| P3 · Practical exam | 20% | Week 7 · Oct 16 · in-class, no AI |
| P4 · Experimental Project | 15% | Weeks 8–12 · immersive, choice of direction |
| P5 · Final Project | 40% | Weeks 12–15 · exhibited in the VP Studio |
Scale: 95–100 A · 89–94 A/B · 84–88 B · 77–83 B/C · 72–76 C · 66–71 C/D · 60–65 D · below 60 F. More than three unexcused absences: NF — excused absences are accommodated, see the syllabus. Every production project includes a process log; from Week 9, when the CLI opens, it also records the agent work as an AI development log.
Equipment & software
- No laptop required — every student has a lab PC in CCB-134 that runs Unreal Engine 5.8 and Twinmotion (Epic account required). Your own laptop, or a gaming desktop at home between classes, works too.
- Claude Code and Codex CLI installed (setup in Week 9, when the CLI opens; see the AI Lab).
- Suggested: 1 TB external SSD; flat-response studio headphones.
- UT Adobe Creative Cloud is free for this class: creativecloud.adobe.com with your UT email.
University policies (Title IX, ADA, attendance, academic integrity, campus closure, disruption) are stated in full on the official syllabus in SpartanLearn / Canvas — that document governs.
Semester Map
Fourteen Friday studio sessions (Sep 4 – Dec 11, no class Nov 27) plus a finals-week exhibition, in four phases: learn the engine, make it interactive, take it into space, then ship the final. Weeks 1–8 are hands-on only; the amber notes from Week 9 mark each week's agent lab.
Phase 1 — The 1:1 World · Weeks 1–3
The 1:1 world: Cesium and the digital twin
- Borges' map at 1:1 with the empire it describes — and what changes when we actually build one. Simulation, hyperreality, and the digital twin.
- Install party: Epic account, UE 5.8, a Cesium ion account and token, the Cesium for Unreal plugin.
- Stream world terrain, georeference to real coordinates in decimal degrees, and add photogrammetric buildings with your own API key.
- Write down where you are: coordinates, sources and attribution, saved in your project folder.
Camera craft: Sequencer and the flyover
- The camera is the argument. Why sudden turns destroy a viewer's mental model of terrain, and why "no rapid movement" is a rule rather than a preference.
- Focal length as a claim about distance: wide exaggerates depth and invites you in, long compresses ridges into unreachable layers.
- Cine Camera Actor, a ninety-second Level Sequence, few keys far apart, everything eased.
- Render properly through Movie Render Queue, then encode with FFMPEG — not a viewport capture.
Blockout & true scale
- Grayboxing a space; 1:1 measurement between real and virtual environments.
- Reference gathering; measuring a real room and rebuilding it.
- Crit: Digital Twin Flyover (Project 1) due; Project 2 launches.
Phase 2 — Building at True Scale · Weeks 4–6
The Twinmotion round-trip: Datasmith into Unreal
- Twinmotion → UE via Datasmith; CAD/BIM import; archviz conventions.
- Cameras, composition, and still/flythrough rendering with Accumulation DoF.
Materials & surfaces
- PBR materials and Substrate basics; material instances.
- Megalights (now production-ready) vs. Lumen vs. Lumen Lite — when to use which.
Lighting & Megalights
- Megalights, Lumen and Lumen Lite — what each buys and what each trades away.
- A base rig that reads as a specific hour, then practicals motivated by fixtures in the model.
- Measure the cost of a lighting pass instead of admiring it.
- Crit: Architectural Visualization (Project 2) due.
Phase 3 — Interactivity & the Experimental · Weeks 7–10
Guided training practical exam
- In-class, hands-on: build a small scene to spec — modeling, materials, lighting, packaging.
- No AI agents during the exam: this is the skill floor the rest of the course stands on.
- After the break: introduction to Blueprints — nodes, events, variables.
Environments, terrain & procedural generation
- Landscape tools vs. experimental Mesh Terrain (overhangs, tunnels, floating islands).
- Foliage, PCG basics, atmosphere and skies.
- Project 4 (Experimental Project) launches — the last hands-only week.
Weeks 1–8 are hands-on only — you build in Unreal yourself, and the AI TA answers questions but nothing else. From Week 9 the agent lab returns to the third block and runs to the end of term. You cannot judge whether an agent did good work until you can do the work.
Blueprints I: making the world respond
- Triggers, timelines, doors, pickups, simple UI widgets.
- Reading a Blueprint you didn't write — the essential AI-era skill.
Blueprints II: interfaces & dispatchers
- Blueprint communication, casting, interfaces; when a project wants C++ instead.
Phase 4 — Into Space, and the Final · Weeks 11–15
VR I: deploying to headsets
- OpenXR, project settings for VR, performance budgets, packaging for Meta Quest and HTC Vive.
- Comfort and locomotion: teleport vs. smooth, vignettes, what makes people sick.
VR II: hands, bodies, presence
- Motion controllers, grabbing, spatial UI; audio as a spatial material.
- MetaHumans and crowd-scale MetaHuman Collections for populated spaces.
- Crit: Experimental Project (Project 4) due; final project pitch.
Production sprint I
- Studio time on finals with instructor desk-crits; sequencer and cinematics for documentation.
Production sprint II & playtesting, in the Volume
- Structured playtests in headsets and on the wall; triage what feedback to act on.
- Packaging, install, and exhibition logistics.
Exhibition & final crit
- Monday Dec 14, 8:30–10:30 am — our final exam slot. Exhibition in the Virtual Production Studio and headsets; invited guests.
- AI development log and process documentation submitted with the build.
Projects
Five projects. The first puts you on the real earth; the last lets you build a world of your own. Every project ships with a process log. From Week 9, when the CLI opens, that log also records the agent work — what you asked for, what it did, and what you had to fix or decided to reject.
01 · Digital Twin Flyover
10%Weeks 1–3You begin with the whole earth. Stream real terrain and photogrammetric buildings at true coordinates with Cesium for Unreal, choose a landscape anywhere on the planet, and make a ninety-second cinematic flyover of it.
The camera is as much the assignment as the terrain: sequence it so the move is legible and unhurried — no sudden turns, no rapid movement. A flyover that lurches tells the viewer nothing about the place.
Deliverables: a georeferenced level · the ninety-second render · your coordinates documented · visible attribution for every source you streamed.
02 · Architectural Visualization
15%Weeks 3–6Having flown over the real world, you build inside it. A designed space at strict 1:1 — measured, blocked out, brought through Twinmotion into Unreal via Datasmith, materialled honestly and lit to a specific hour.
Sited, not floating: the building stands at true coordinates with real neighbors and a matched sun. A before/after pair and one thirty-second approach shot in context.
03 · Practical exam
20%Week 7Stations, timed, hands-only: no agents, no chat, no AI TA, no internet. Includes a geo station — georeference, anchor, set the sun to the exam hour — on Cesium World Terrain and OSM Buildings, so nothing depends on a quota under time pressure.
Weeks 1–8 are hands-on precisely so this exam tests what you have actually been doing.
04 · Experimental Project
15%Weeks 8–12The open one. An immersive or experimental work that is yours — spatial self-expression rather than a specification met. A VR piece, a projected or LED-wall installation, a data-driven landscape, a VP Studio collaboration, or something the brief did not anticipate.
It launches the week before the CLI opens, deliberately: you scope it by hand, then build it with agents available from Week 9 and log what they did.
05 · Final Project
40%Weeks 12–15Also a choice, at the scale the VP Studio allows. Most students build an immersive world. You can build for a headset — Vive and Quest units are available, you buy nothing. Or you can make a snowglobe: design one, and let the player navigate the inside of the sphere. Screen-playable, and a sharper design problem than it sounds — no horizon to hide behind, every direction a composed view.
The Week 15 pre-show audit is mandatory: attribution visible, data-source credits present, ion tokens valid through the run.
AI Lab
The agents arrive in Week 9, not Week 1. Weeks 1–8 are hands-on only — you build in Blender and Unreal yourself, because you cannot judge whether an agent did good work until you can do the work. From Week 9: This course treats AI coding agents the way a studio treats any powerful tool: you are the director, the agent is a fast but overconfident assistant, and nothing ships that you can't explain. Here's the toolchain, the setup, and the rules. And this site has its own AI TA — the amber button in the corner of every course page — tuned to each week's material for questions between studio sessions, all fifteen weeks of the term.
The toolchain
Claude Code
Anthropic's terminal-based coding agent. Reads your project, edits files, runs builds. With Epic's official UE skills plugin it understands engine conventions; with MCP it can operate the editor itself.
OpenAI Codex CLI
OpenAI's counterpart agent. We use both deliberately — comparing how two different agents attack the same Unreal problem is itself a course exercise (Week 9).
Unreal MCP (built into 5.8)
An MCP server inside the Unreal Editor exposing 30+ toolsets — actors, Blueprints, materials, Niagara, Sequencer, automation tests — so agents can act in the editor, not just on files.
Epic's UE skills for Claude Code
An official Epic plugin that gives Claude Code built-in knowledge of UE workflows and documentation lookup, so its answers cite engine reality instead of guessing.
Setup (Week 9, step by step)
- Install the agents
Claude Code (
npm install -g @anthropic-ai/claude-code) and Codex CLI per current docs. Sign in with the accounts provided/discussed in class. - Enable Unreal MCP in the editor
In UE 5.8:
Edit > Plugins, search "Unreal MCP", enable it (Toolset Registry enables automatically), restart. Start the server from the console:ModelContextProtocol.StartServer. - Connect Claude Code to the editor
Add the local MCP endpoint to Claude Code and install Epic's unreal-engine-skills-for-claude-code plugin so the agent has engine-aware skills.
- Snapshot before, compare after
Before any AI session: copy your project folder aside. After: compare what changed against the copy. That before/after pair is your ground truth for the AI log.
- Smoke test
Ask the agent to list the actors in your level and describe the project structure. If it can see your world, you're wired up.
Prompt patterns that work in Unreal
- Look up before you build. Have the agent use its class/docs lookup (inheritance, properties, functions) before writing Blueprint or C++ logic — UE's API is too big to trust from memory.
- Small, verifiable steps. "Add a trigger volume that opens this door" beats "build my level's interactivity." Compile and test after every change.
- Spec first. Write the interaction as three plain-English sentences before prompting. If you can't spec it, the agent can't build it — it will just build something.
- Make it explain. After the agent edits a Blueprint, have it walk you through the graph. Then verify in the editor that the explanation matches the nodes.
- Distrust performance claims. Agents assert; profilers measure.
stat unit,stat gpu, and Unreal Insights settle arguments.
Weeks 1–8 — hands only. No agents in the work: no Claude Code, no Codex, no generative tools producing geometry, textures, rigs or code you submit. Ask the AI TA anything you like about how something works; do the work yourself.
Week 7 — nothing at all. No agents, no chat, no AI TA, no internet. The practical exam is unaided by design.
Weeks 9–15 — allowed and expected, and logged. Use Claude Code and Codex on production work, with every session recorded in your AI development log: prompt, what the agent did, what you kept, fixed or rejected.
Required all term: you must be able to explain every asset, material, animation state and line of code in your submitted build, live, at crit. "The AI did it" is a starting point, not an answer.
Never allowed: presenting agent output as hand-authored; using AI to fabricate playtest notes, references or reflection writing; agents on Week 1–8 submitted work or during the Week 7 exam.
Under the University's Undergraduate Academic Integrity Policy, instructors may identify conduct that prevents fair evaluation of your learning as a violation. In this course that includes agent use in Weeks 1–8, any AI use during the Week 7 exam, and undocumented AI use from Week 9 on.
A running document (Markdown in your project folder is ideal) with one entry per working session: date, goal, the prompts that mattered, what the agent produced, what you changed and why, and one thing you learned about directing it. Five honest minutes after each session. It is worth 15% of every project grade. Before Week 9 you keep the same habit as a process log, minus the agent: what you tried, what failed, what you decided.
Blender in Five Days
Five short sessions to get Blender under your hands before this course needs it — forty-five to seventy minutes each, self-paced, ungraded, in any week you like. Unreal is where this course lives, but the ground you sculpt and the assets you build start in Blender, and every one of them has to survive the trip. These five days are that trip.
This sits inside the hands-only stretch of the course, so: no Claude Code, no Codex, no generative tools making your geometry, rigs or animation. The AI TA is open the whole time for questions — asking it why did my terrain import with stair-steps is help, and it is encouraged. Having it build the mesh is not.
From Week 9 you will be directing agents, and you cannot judge whether an agent did good work until you can do the work yourself. These five days are part of what makes that judgement possible.
The two strands
Ground
The unit contract with Unreal, then landforms: what belongs in the engine's Landscape system, what has to be a modeled mesh, and why a heightfield physically cannot make a cave.
Character & pipeline
A rig the engine will accept, a walk cycle that loops, and the export pipeline that decides whether any of it arrives intact.
Days build on each other: Day 5 needs the terrain from Day 2 and the walk from Day 4.
Never opened Blender? Start with Day 0 — an hour on the interface and the words the other days use. Day 1 is required reading for everyone, and it is forty-five minutes. Week 3 asks you to blockout at true scale and Week 4 round-trips assets through Datasmith; both assume you know how size survives an export. Day 2 is the upstream half of Week 8's terrain material. Day 5 applies to every asset you will ever move into the engine.
Days 3 and 4 are optional unless your final has a character in it, or you want to understand what Week 12 is doing when it puts a body into VR. This course has no MetaHuman work and Projects 1–3 are environments.
Ground
The interface, and the words for it
- For anyone who has never opened Blender. The window named part by part, moving around, Object mode versus Edit mode, G/R/S and the N panel, the Add menu and the 3D cursor, what a modifier is, and where the Adjust Last Operation panel hides. Every step on the later days assumes these.
- Two Preferences settings for PC laptops and trackpads (Emulate Numpad, Emulate 3 Button Mouse), a glossary of the words the later days use without explaining, and a ten-minute drill to prove you are ready.
- Screenshots throughout — and from here on, every day page has them at the steps that are hardest to follow from text alone.
The port contract
- Blender measures in meters; Unreal measures in centimeters. One Blender meter must arrive as 100 Unreal units. That sentence explains most "why is my model enormous" problems you will have this term.
- Build a cube exactly 2 m, Ctrl+A → All Transforms, export, import, and measure it in Unreal — it should read 200. Then break it deliberately and watch the failure appear.
- Read the N panel properly: what Dimensions, Scale and Rotation are each claiming, and which state you want before you export.
Landscape, end to end
- Two kinds of ground, and the distinction governs everything. A heightfield stores one height per point, so it physically cannot make a cave mouth or an overhang. Those have to be meshes. That is the whole answer to "why won't the terrain tool sculpt this".
- Sculpt a landform: proportional editing for the big shapes, then Draw, Clay Strips and Smooth — and judge it from eye height, not from orbit.
- Both routes into Unreal: a 16-bit heightmap into the Landscape system (8-bit terraces visibly), and mesh export with a UCX_ collision mesh for the overhangs.
Character & pipeline
A rig that survives the trip
- Rigify — which ships with Blender but is off by default, so you will not find it in a menu until you enable it. Fit the metarig, generate, bind with automatic weights.
- Fitting the metarig is the job. A bad fit cannot be rescued by weight painting afterwards.
- The four things Unreal requires: a root bone at the origin separate from the pelvis, deform bones only, applied transforms on the armature, and .L / .R naming left intact.
Making it move
- A walk cycle is four core poses — contact, down, pass, up — not thirty. Block all four before polishing any of them; the commonest failure is spending forty minutes on pose one.
- In the graph editor, spacing between keys is timing. Evenly spaced reads as mechanical; bunched at the extremes reads as mass.
- Loop it properly: first and last frames identical, played one frame short, or the doubled frame reads as a stutter every cycle.
The export pipeline, and the round trip
- Root motion or in-place — decided before you animate, one per clip, consistently. Mixing them is what makes a character moonwalk or run away from their own collision capsule.
- Name clips like assets, match the frame rate to the engine, bake on export, and import the character first so every animation lands on one skeleton rather than accidentally creating a second.
- Then prove it: Day 2's terrain and Day 4's character in one scene, correct scale, feet on the ground, loop clean.
Each of the day pages carries its own tutor — a separate one from the course AI TA, briefed on Blender specifically, on this path, and on the shelf below. It knows which day you are reading, so you can ask “why did step 4 do that” without explaining the context first.
It explains and diagnoses; it does not do the work. Ask it why your normal map is inverted and it will walk you through it. Ask it to write a script that generates your terrain and it will decline — that is the whole point of these days. You will need the class password the first time you open it.
Where to learn more
A short, deliberately curated shelf — and the same one the Blender TA draws on, so what it recommends is what is listed here. Three layers, and it is worth knowing which kind you are opening before you start.
Reference
What a tool is, and what its options mean. Go here to look something up, not to learn an approach.
| Resource | Level | Best used for |
|---|---|---|
| Blender Reference Manual | all | The canonical documentation and the backbone of this taxonomy. Send students here for what a tool IS and what its options mean — not for how to approach a project. |
| Blender Support hub | all | Blender's own starting point, pointing at the manual, community tutorials and Blender Studio. |
| Blender Tutorials (official index) | beginner | The officially curated tutorial entry point. |
Tutorials
How a workflow actually goes, watched over someone’s shoulder. Tagged by teaching style, because the right one depends on where you are stuck.
| Resource | Level | Best used for |
|---|---|---|
| Blender Guru search by name | beginner | Step-by-step foundational projects; the Donut series is the usual first thing a beginner is pointed at. Best for building confidence with a finished result. |
| Grant Abbitt search by name | beginner | Approachable, patient beginner instruction. The best recommendation for a student who found another tutorial too fast. |
| CG Geek search by name | beginner | Broad practical Blender workflows across a wide range of subjects. |
| Ryan King Art search by name | intermediate | Short, focused single-technique tutorials. Best when a student needs one specific thing rather than a project. |
| Polygon Runway search by name | intermediate | Stylised and illustrative work at an intermediate level. Good for a student who can drive Blender but is asking how to make things look designed. |
Courses
A sequenced path with an order someone thought about. Worth it if a playlist leaves you unsure what to do next.
| Resource | Level | Best used for |
|---|---|---|
| Blender Studio | intermediate · paid | Production-oriented training and real production assets, from the Blender organization itself. The closest thing to seeing how professionals actually work. |
| Class Central Blender roundup | all | A meta-curation layer for comparing courses across providers. Use it to vet, not to learn. |
| Coursera — Blender for Beginners Specialization | beginner · paid | A sequenced beginner curriculum for students who want structure rather than a playlist. |
Several of the tutorial creators are named without a link, on purpose. Channel URLs move, and a dead link that looks authoritative wastes more of your time than no link at all — search the name instead. The Blender TA is instructed never to invent a URL either; if it does not have one, it will tell you what to search for.
The topic buckets here mirror the official manual's own structure rather than our course weeks. That is deliberate: the manual gets reorganized rarely, so this shelf stays usable long after this semester.
What these five days are not
Box modeling, UV layout, texturing, baking, Geometry Nodes, Cycles rendering, physics and Grease Pencil are all outside this path. That is deliberate: what is here is the part of Blender that touches Unreal, which is the part this course runs on. If you want the broader modeling and texturing track, that is FMX 320.
Headroom
In audio, headroom is the margin between your loudest signal and the ceiling. Run out of it and the sound doesn't get louder — it clips. Your attention works exactly the same way. This is a note on cognitive load: what it is, how this course is built around it, why AI tools raise it rather than lower it, and how to protect your margin while carrying four courses at once.
Working memory is small, and it does not expand. That is the founding observation of cognitive load theory, and everything on this page follows from it. When a project asks you to hold creative judgement, software navigation, technical settings, file management and a deadline in your head at once, the first thing to fail is not your talent — it's your margin.
Professionals are not people who push harder against the ceiling. They are people who engineer their signal chain so they rarely hit it. That's a learnable skill, this course deliberately trains it, and it will matter to your career at least as much as anything you build here.
Three kinds of load
Cognitive load theory splits mental effort into three kinds. Learning to tell them apart in the middle of a work session is the whole trick.
The task itself
The necessary complexity of the work — editing a scene while holding continuity, pacing and intent; a shoulder that has to deform believably. You can't remove it. You can only shrink it with prior knowledge: whatever becomes automatic stops costing working memory. That is what practice is for.
The friction
Effort the work never asked for — hunting for the brief, decoding a vague spec, un-versioned files, redoing a lost export, guessing what the rubric wants. It teaches nothing and clips you anyway. Kill it with systems: naming conventions, one folder template, one home per project.
The learning
The productive effort that builds durable skill — comparing two versions, explaining why the cut works, applying the principle to the next piece. This is the load worth paying for, and it's exactly the load that friction crowds out first.
The goal is never an easy course. The goal is to spend your ceiling on the work that compounds.
This course is built around your headroom
Several things about this course that can read as ritual are load engineering, and you'll get more out of them knowing why they're there.
- Weeks 1–8 are hands-only because chunking is real. A skill you've practiced until it's automatic stops occupying working memory — which is what frees you, later, to think about the work instead of the tool. The Week 7 practical isn't a hazing ritual; it's a headroom reading. It measures what you can do without looking anything up.
- The recurring structures are friction removal. The same session rhythm every week, one submission convention, the same crit format: you learn each format once, and from then on your attention goes to the work. That's extraneous load being deleted on your behalf.
- Demos are segmented and models are narrated — short named steps, worked examples with the decisions explained — because research is unambiguous that learners absorb more, with less effort, when complexity arrives in pieces they can process.
- The process log is external memory. Paper is cheap; working memory isn't. A decision written down is a decision you can stop carrying. By Week 9 the same habit becomes your AI development log, and it's graded because it's the skill.
The AI part: load doesn't drop — it converts
It is tempting to expect the agents to arrive in Week 9 and lower the effort. They won't. They convert it. Before Week 9, the cost of your work is generation — slow, but you understand every asset because you made it. From Week 9 the cost moves to verification and orchestration: reviewing work you didn't do, holding your intent steady across ten fast iterations, deciding what to keep. Generation is bounded by your hands. Supervision is bounded by your judgement — and an agent can outrun it.
That's why this course calls the agent a fast but overconfident assistant. Every output you accept without verifying is borrowed load — it comes back with interest at crit, in the exam room, or in the build that won't package the night before the showcase.
And the career part, plainly: everyone you'll compete with will have the same agents you do. The day everyone can generate, production stops being the scarce skill. What stays scarce is a head that stays clear while directing ten times the output — scoping, verifying, explaining. In a tighter market that difference is the hire. Your AI development log isn't paperwork; it's the portfolio artifact that proves you can supervise the machine accountably.
Four courses, one ceiling
You're carrying this course alongside three others, plus a life. The ceiling doesn't care that the loads come from different syllabi — it's one working memory. Six moves that protect it:
Batch, don't shred
Context-switching is the most expensive thing you do all week. One deep block per course beats four shredded evenings, because every switch pays a re-loading tax and leaves half-finished thoughts resident in memory.
Write it down
A capture list is external memory: ideas, due dates, “fix this later.” If it's written, it can stop occupying you. The open loop you're trying to remember is the one eating your margin during the work.
Close loops
End every session with one line of state: “next — reweight the shoulder.” Future-you boots in seconds instead of twenty minutes. Same rule as the agent log: write the state down before you move on.
One home per course
Every course gets one predictable place for briefs, files and deadlines. Five minutes spent hunting is pure extraneous load; spend the five minutes once, on the folder template.
Scope is load management
Ambition is graded alongside finish here for a reason. Cutting scope early isn't defeat — it's the production judgement studios actually pay for, and it's how a final ships polished instead of ambitious and broken.
Audit weekly
Ten minutes, all four courses plus life, everything on one page. Tag each item: generate, verify, or coordinate. You cannot route a load you've never seen written down — and most people have never once seen theirs.
Four questions, thirty seconds, any course: Does it read without me explaining it? Can I explain every part of it at crit? Is it in the required format, named the required way? Did I write down the one thing I learned? If all four are yes, submit and close the loop.
Where this comes from
Cognitive load theory begins with John Sweller's work in the 1980s and is one of the most replicated bodies of research in instructional design. If you want the sources behind this page:
- Sweller — Cognitive load theory, educational research, and instructional design (Instructional Science)
- van Gog et al. — advances on worked examples, animations, and load measurement (Educational Psychology Review)
- NSW CESE — Cognitive load theory: research that teachers really need to understand (the best short practical guide)
- Spanjers et al. — segmentation of worked examples (why the demos come in named pieces)
- Every Learner Everywhere — cognitive load and equity in higher ed
The application to media production, four-course semesters and AI supervision is mine. Argue with me about it at crit.
Resources
The bookmarks that matter. Everything here is current for UE 5.8 as of summer 2026.
Engine & tools
- Unreal Engine — download via the Epic Games Launcher.
- UE 5.8 release notes — Megalights, Lumen Lite, Mesh Terrain, MetaHuman Collections, Toon Shader.
- Twinmotion — archviz pipeline partner for Project 2.
- Epic Developer Community — Learning — official guided courses used for the Week 7 training exam.
AI agents & MCP
- Unreal MCP in the Unreal Editor — official docs for the in-editor MCP server.
- Epic's UE skills plugin for Claude Code — official engine-aware skills.
- Claude Code documentation — install, MCP configuration, workflows.
- mcp-unreal (community) — headless builds/tests, Blueprint editing, API lookup; works with Claude Code and Codex.
XR & virtual production
- Meta Quest developer documentation — Quest packaging, performance budgets, comfort guidelines.
- OpenXR in UE — the one API path we target for both Quest and Vive.
- nDisplay & ICVFX docs on Epic Developer Community — the VP Studio's software backbone.
State-of-the-art reading list
The sources behind each week's SOTA segment. Claims fact-checked July 2026; market figures are analyst estimates and Epic performance numbers are vendor targets.
- Kerbl et al., “3D Gaussian Splatting for Real-Time Radiance Field Rendering” (SIGGRAPH 2023) — the course's landmark paper.
- Omdia: XR headwear outlook (June 2026) and Counterpoint quarterly XR tracker — the market reality.
- Epic: Unreal MCP in the editor — first-party docs for the agent workflow.
- Futuresource: LED volumes in virtual production — industry scale (2024 vintage).
- IEEE VR 2025 best papers — human-factors research entry points.
- MetaHuman Animator markerless mocap (free) — phone-video performance capture.
University
- SpartanLearn (Canvas) — announcements, submissions, the governing syllabus.
- UT Adobe Creative Cloud — creativecloud.adobe.com with your UT email.
- Student Accessibility Services — (813) 257-5757 · accessibility.services@ut.edu.