Jiayao LANG's Portfolio - Overview

90°N - A Psychological Healing Game

I. Overview & Galleries

• Completion Time: June 2024 – July 2025

• Project Type: Virtual Reality Computer Game (with an additional first person keyboard edition)

• Role: Solo Developer (Design, Art, Programming, Technical Implementation)

• Tools Used:

  Software	Usage
  Unreal Engine	Core Development, Blueprint Logic, VR Interaction
  Maya	Modeling and Rigging weapons for first person
  Substance Painter	Texturing Weapons, and texturing special landscape textures
  Quixel Megascans	Natural env assets and surface textures
  Photoshop & Illustrator	Crafting VFX textures and UI elements
  XMind	Concept collection
  Fooocus & CosyVoice	Generating voices of narrator, and concept illustrations
  GeoGebra	Debugging math algorithms for dynamic VFX shaders
  Cascadeur	Animating first person animations (arms and interactions)
  Gaea	Terrain sculpting and landscape heightmap generation

  Note
  

  Fooocus and CosyVoice are open‑source AIGC software tools.

• Game Screenshots:

  

  

  

  Note
  

  More detailed screenshots and gameplay footage can be found on ArtStation.

   

II. Description

1. Background & Concept

90°N is designed as a healing‑oriented interactive experience, using the quiet vastness of the Arctic as a setting for emotional restoration. Through spacious environments, interesting interactions, the game offers players a calm refuge from the intensity of everyday life.

Goals & Themes

• Emotional Expression Through Interactivity: Explore the unique potential of games to communicate feelings and perspectives through player.

• Emotional Calm & Reflection: Use slow-paced, travel‑oriented gameplay to create a meditative experience that helps players decompress.

• Stress Relief and Emotional Well‑Being: Provide a calming digital space that supports mental decompression.

Target Audience / Context

• People seeking emotional relaxation: Looking for a calming digital space to unwind, reflect, and temporarily step away from the fast pace of daily life.

• Casual players looking for accessible, non‑stressful gameplay: People who may not identify as “gamers” but enjoy intuitive, low‑pressure experiences that offer beauty, curiosity, and emotional comfort.

2. Research & Analysis

Ideation & Thought Development

The early concept development for 90°N began with a video exploring the emotional and ecological dimensions of polar travel. The ideation process focused on simulating the psychological healing of Arctic exploration while offering meaningful interaction.

• Global Warming & Glacier Preservation: The concept emphasized the urgency of documenting preserving the glaciers, which are rapidly vanishing due to climate change.

• Mental Healing & Emotional Resonance: The game’s slow pacing, aurora scenes, and quiet exploration were designed to evoke inner calm, joy, and reflection.

Narrative & Scene Prototypes

• A tent camp serves as the player’s sanctuary, player could switching between day and night to reveal different emotional and visual states of the environment.

• A nighttime polar journey across vast ice fields leads to a breathtaking aurora event.

• A helpless polar bear drifting on a shrinking ice floe, quietly illustrating the consequences of global warming.

3. Key Design & Production Process

The production pipeline for 90°N followed a clear workflow that combined environment creation, interaction design, VR systems, and visual effects. Each stage used specialized tools to keep the process efficient and easy to iterate.

1. Early UI & Level Planning

   To establish the structure of user experience, I first created a set of UI sketches and layout references using Illustrator. These early drafts helped define navigation flow, interaction points, and the emotional pacing of each scene.

   

2. Landscape Creation with Gaea & Unreal Engine

   The Arctic terrain was built using Gaea, where I sculpted the base landscape, snow ridges, and glacier formations. The heightmaps were then exported into Unreal Engine for further refinement. Inside Unreal, I used Nanite Displacement, Layer Blending, and procedural generation technology to achieve natural surface variation.

   

    

3. Environment Assembly & Visual Atmosphere

   Using Megascans assets, I populated the world with rocks, ice surfaces, and natural debris using PGC workflow. Material shaders were customized to simulate snow covering and blending, subsurface scattering on floating glaciers, anti-tiling.

   

    

   And an early playable build was created to test scale, lighting, and traversal.

   

4. FPV Animation & Rigging

   All first person models and animations — such as camera with equipping animation, were created in Maya, and textured in Substance Painter. And using the model without arms for VR gameplay.

   

    

5. Interaction Systems & Inventory Design

   A custom inventory system was developed for both the keyboard version and the VR version. The VR inventory opens with a scanning‑dot VFX inspired by the game Death Stranding.

   

    

6. Timelapse Tools

   A VR time‑manipulation tool was created to let players switch between day and night. This system uses particle‑based holographic elements with a little electronic disturbing visual effect, showing the sun, moon, and hour markers . For keyboard experience, replace it with a modern appearance clock, player can drag the circular progress bar and set the duration of time passing.

   

7. VR Gallery & Memory System

   Players can capture photos during exploration. These photos appear as floating holographic squares inside a VR gallery space, allowing players to revisit their journey. The photos taken by the player will be save to the local disk, so the gallery wont forget player's experience.

   

    

8. Special Interaction: World‑Traveling Orb

   A unique VR gameplay allows players to grab a glowing “world sphere” similar to a divination orb. VR player could dive in to the world by taking closer to the ball. The faster the player swings the ball, the flatter it will become, which is based on vertex offset shader.

   

    

    

4. Final Outcome

High-Fidelity Results

• Project Online Portfolio: ArtStation (including videos below)

• Project Online Video:

• First-Person Edition

• Virtutal-Reality Edition

Steel Defender - A Multi-Platform Multiplayer Game

I. Overview & Galleries

• Completion Time: December 2024 – May 2025

• Role: Solo Developer (Design, Art, Programming, Technical Implementation)

• Tools Used:

  Software	Usage
  Unity Engine	Core Development, VR Interaction
  Maya	Modeling and Rigging weapons
  Jetbarins Rider	Writing C# Scripts
  Substance Painter	Texturing Weapons, and texturing special landscape textures
  Quixel Megascans	Natural env assets and surface textures
  Photoshop & Illustrator	Crafting VFX textures
  CosyVoice	Generating voices of narrator
  GeoGebra	Debugging math algorithms for dynamic VFX shaders
  Gaea	Terrain sculpting and landscape heightmap generation

  Note
  

  CosyVoice are open‑source AIGC software tool.

• Game Screenshots:

  

  

  

  Note
  

  More detailed screenshots and gameplay footage can be found on ArtStation.

   

II. Description

1. Background & Concept

Steel Defender is a cross‑platform multiplayer shooter game. VR players and keyboard‑mouse players can play together in the same world while taking on different roles. The game features both a first‑person mode and a VR mode, and its detailed environments create a strong sense of visual impact.

Goals & Themes

• Explore Cross‑Platform Gameplay Design: Define clear gameplay roles and interactions for both VR players and keyboard‑mouse players.

• Create High‑Quality Visual and Interactive Systems: Use Shader Graph to build convincing visual effects and apply the XR Interaction Toolkit to design rich, responsive weapon interactions.

• Completely Gameplay Flow: A thrilling and fast‑paced gameplay experience with engaging cutscenes that enrich the overall gameplay experience.

2. Research & Analysis

Ideation & Thought Development

The early concept development for this project began with observing how players increasingly move between different devices in their daily lives. This inspired the idea of creating a shared game world where PC, mobile, console, and VR players can interact smoothly and enjoy the same experience.

• Representative Cross‑Platform Game: Games like Phasmophobia demonstrated how interesting interactions and multi‑platform connectivity can bring players together across PC and VR.

• Rich VR Interaction: VR adds natural hand interactions and immersive weapon handling, giving VR players unique ways to engage while staying connected with others.

3. Key Design & Production Process

The production workflow for this project followed a structured pipeline that combined VR interaction design, multiplayer networking, environment creation, visual effects, and asset production.

1. VR Interaction & Multiplayer Framework

   The early prototype focused on building a multiplayer framework that supports both VR players and keyboard‑mouse players in the same session. This stage ensures that VR’s physical interactions and traditional keyboard controls could coexist smoothly.

   

2. Floating Enemy AI Prototype

   A helicopter enemy was created, and test floating AI behavior. Its vertical motion used a sine‑based height jitter, allowing the prototype to simulate hovering, strafing, and attacking block targets.

   

    

3. Pistol Asset Creation

   The VR pistol was modeled in Maya, textured in Substance Painter, and then rigged for animation. Inspired by P102 in the game HELLDIVERS2.

   

    

   By separating the hand from the weapon and letting both share the same hand skeleton, the workflow becomes clearer, more modular, and easier to extend to other weapons, or just animate the hand gesture.

   

4. Ammo Indicator VFX & Hologram Visual Effect

   Inspired by Half‑Life: Alyx, a holographic ammo indicator was designed in Photoshop, implemented in Unity, giving the VR pistol a more futuristic VR‑friendly interaction cue.

   

    

5. Level Design Prototyping

   The level prototype explored how the helicopter enemy could guide the player through the environment. Megascans assets were used for quick iteration, and a single stone model was varied through Shader Graph to maintain performance while preserving visual diversity.

   

6. Environmental Shaders & Atmosphere

   Several custom shaders were developed to enhance environmental realism: snow covering, lightweight volumetric fog using intersecting planes, procedural grass variation where color and size were sampled from world position to create natural randomness, and etc.

   

    

7. Landscape Creation with Gaea

   The terrain was generated in Gaea, using erosion nodes and height‑based masks to sculpt a natural landscape. The skybox is rendered using a MatCap‑based approach, allowing it to produce a striking visual effect that enhances the overall atmosphere.

   

    

    

    

4. Final Outcome

High-Fidelity Results

• Online Portfolio: ArtStation (including videos below)

• Online Video: YouTube

Digital Teacher Xiao E

I. Overview & Galleries

• Completion Time: February 2024 – April 2025

• Role: Solo Developer (Design, Art, Programming, Technical Implementation)

• Tools Used:

  Software	Usage
  Unreal Engine	Core Development
  Maya	Modelling, Rigging, Animating
  Jetbarins Rider	Writing C++ Plugins
  Marvelous Designer	Designing clothes
  Houdini	Crafting Grooms
  Photoshop & Fooocus	Crafting backgrounds, balance compositions, clean up photos
  Reality Capture	3D Scanning human face
  MetaHuman Creator	Creating human model and offering expression technology
  Audio2Face	Generating runtime facial expressions for MetaHuman
  Anything LLM	Offering enhanced functions for LLM service, such as RAG database
  Qwen GPT 1.5B	Offering local LLM service
  ChatGPT	Offering online LLM service
  Sherpa-NCNN	Offering speech-to-text service
  edge-tts	Offering text-to-speech service

  Note
  

  Fooocus(AIGC), Sherpa-NCNN, edge-tts are open‑source software tool.

   

• Screenshots:

  

  Note
  

  More detailed screenshots and footage can be found on ArtStation.

   

II. Description

1. Background & Concept

Xiao E is a digital‑twin teaching assistant powered by 3D scanning and AI speech recognition. Xiao E enhances teacher-student AI interaction, offers immersive guidance, answers campus questions, and supports digital campus services.

2. Research & Analysis

Ideation & Thought Development

The early concept for this project began with the idea of creating an intelligent Q&A assistant that could appear on a semi‑transparent display using MetaHuman technology. I wanted this assistant to communicate naturally with students and feel believable enough to be accepted as part of their daily campus experience.

• AI‑Driven Interaction: The assistant needed to answer student questions through an AI‑powered dialogue system.

• Realistic Appearance: MetaHuman technology was chosen to give the assistant a realistic appearance that students could trust and engage with.

• Local Knowledge Base: A built‑in local database ensures the assistant can deliver customized information and speak the content I design for specific educational scenarios.

3. Key Design & Production Process

The production workflow for this project followed a structured pipeline that combined VR interaction design, multiplayer networking, environment creation, visual effects, and asset production.

1. Head Model & Base Asset Creation

   The workflow began with a scanned real‑person face model, which served as the foundation for creating a MetaHuman facial asset with identity‑tracking features in Unreal Editor.

   For privacy reasons, I cannot show the RealityCapture screenshots, the main work there involved cleaning the photos and removing model noises. The face model was then further refined into an animated style in Maya.

   Inside MetaHuman Creator, the character initially appears without hair, and refinement are completed in later stages.

   

2. Hair Grooming in Houdini

   Hair was crafted using Houdini’s groom system, allowing precise control over powerful hair guidelines. It can also export Hair Card easily. This ensured the final appearance would maintain a natural and well performance in real‑time rendering.

   

3. Cloth Creation in MD & Rigging in Maya

   Clothing was designed in Marvelous Designer for realistic cloth simulation, then exported to Maya for binding and skin‑weight adjustments.

   

   Note
   

   The clothing shown in the picture is from an earlier time. I also created a digital human before Xiao E.

4. Assembly in Unreal Engine

   The character’s cloth, body, and hair were composited within Unreal Engine’s MetaHuman workflow. In this workflow, all the component skeletons have to share the same skeleton so they can act in same pose or animation. The torso was omitted for performance optimization while maintaining full visual fidelity for visible regions. For materials, the glasses use Cubemap reflection, and the skin uses flatten normal like an animate style.

   

5. Background Material & Mosaic Transition Effect

   A custom background material was created to transition between two images using a mosaic‑style effect. This was achieved by down‑sampling UV to produce a stylized low resolution and slide from left to right appearance during the transition.

   

   Note
   

   Transition footage can be found on ArtStation.

6. C++ Integration for AI Communication

   C++ modules were implemented to run Sherpa’s binary with parameters, enabling socket‑based communication between UE and Sherpa. Additional C++ logic handled Edge‑TTS on a separate thread, monitoring downloaded voice files for real‑time playback.

   

    

7. Blueprint Logic for ChatGPT Messaging

   Blueprint scripts were built to send JSON messages through HTTP requests to the ChatGPT backend. The system then awaited responses to drive the assistant’s conversational behavior.

   

   Tip
   

   The HTTP sending function in the image is one of the plugins I created, you can obtain it on Fab.

8. Audio2Face Streaming Pipeline for Facial Animation

   When the UE received a voice file from edge-tts, it will run a Python application to push the voice file into Audio2Face’s streaming pipeline. This will generate real‑time facial expressions for the digital human.

   

9. Generate Backgrounds using Fooocus

   Use Fooocus to generate the images for each element, then composition them in Photoshop.

   

   Note
   

   More detailed information about the paintings can be found at here.

4. Final Outcome

High-Fidelity Results

• Online Portfolio: ArtStation (including the video below)

• Online Video: YouTube