A curated, no-fluff collection of Applied AI Researcher (Generative 3D) interview questions with answers, organized by topic. Built for candidates prepping for roles building generative models for 3D content — 3D Generative AI Researcher, Neural Rendering Research Engineer, Text-to-3D/4D Researcher, and Applied Scientist (3D Vision/Graphics) roles — and for interviewers building question banks.
📎 Scope note: This repo sits at the intersection of three fields that a Generative 3D researcher needs fluency in: 3D representations (meshes, point clouds, SDFs, NeRFs, Gaussian Splatting), generative modeling (diffusion, GANs, VAEs, autoregressive/flow models — and how each does or doesn't translate cleanly from 2D images to 3D data), and differentiable/neural rendering (the bridge that lets 2D image supervision train 3D generative models). It's distinct from the Awesome-Perception-Simulation-Engineer-Interview-QA repo in this series, which covers perceiving and simulating the 3D world for robotics/AV, rather than generating novel 3D content from learned models.
Every answer aims to be concise, correct, and interview-ready — the kind of answer that would actually land well in a 45-60 minute technical/research round, not a textbook chapter.
⭐ Star this repo if it helps your prep. PRs adding new questions, fixing answers, or improving explanations are very welcome — see CONTRIBUTING.md.
| # | Topic | Questions | Difficulty Mix |
|---|---|---|---|
| 01 | 🧊 3D Representations Fundamentals | 11 | Easy → Hard |
| 02 | 🧠 Generative Modeling Fundamentals | 11 | Medium → Hard |
| 03 | 🌪️ Diffusion Models for 3D Generation | 10 | Medium → Hard |
| 04 | 📸 Neural & Differentiable Rendering | 10 | Medium → Hard |
| 05 | 🗣️ Text-to-3D & Multimodal Conditioning | 9 | Medium → Hard |
| 06 | 🌌 Neural Fields: NeRF & Gaussian Splatting | 10 | Medium → Hard |
| 07 | 🕸️ Mesh Generation, Topology & Post-Processing | 9 | Medium → Hard |
| 08 | ✨ Texture, Material & Appearance Generation | 9 | Medium → Hard |
| 09 | 🚰 3D Data Pipelines, Datasets & Preprocessing | 9 | Medium |
| 10 | 📏 Evaluation Metrics for Generative 3D | 9 | Medium → Hard |
| 11 | 🏗️ Training Infrastructure & Scaling | 9 | Medium → Hard |
| 12 | 🔬 Research Methodology & Experimentation | 9 | Medium → Hard |
| 13 | 🎭 Scenario-based & Behavioral | 10 | Medium → Hard |
Total: 125 questions in v1, growing with community contributions.
- Cramming for an interview next week? Start with the topic weighted heaviest for your target role (see below), and read the "Follow-up" notes — interviewers almost always dig deeper.
- Deep prep over weeks? Work through every file top to bottom — for the math-heavy topics (diffusion, differentiable rendering), re-derive the key equations yourself; for representation topics, be ready to sketch the tradeoffs on a whiteboard.
- Interviewing candidates? Use these as a base question bank — mix easy/medium/hard per round, and use the scenario/research-methodology questions to gauge genuine research judgment, not just paper recall.
| Role | Prioritize |
|---|---|
| 🚀 3D Generative AI Researcher (generalist) | 3D Representations, Generative Modeling Fundamentals, Diffusion for 3D, Evaluation Metrics |
| 💬 Text-to-3D / Multimodal Researcher | Text-to-3D & Conditioning, Diffusion for 3D, Neural Fields, Research Methodology |
| 📸 Neural Rendering Research Engineer | Neural & Differentiable Rendering, Neural Fields (NeRF/Gaussian Splatting), Training Infra |
| 🗿 3D Asset / Mesh Generation Researcher | Mesh Generation & Topology, Texture/Material Generation, 3D Data Pipelines |
| ⚙️ Research Engineer (infra-leaning) | Training Infrastructure & Scaling, 3D Data Pipelines, Research Methodology |
| 📱 Applied Scientist (product-facing) | Evaluation Metrics, Text-to-3D & Conditioning, Scenario-based & Behavioral |
Awesome-Applied-AI-Researcher-Generative3D-Interview-QA/
├── README.md ← you are here
├── CONTRIBUTING.md
├── LICENSE
└── topics/
├── 01-3d-representations-fundamentals.md
├── 02-generative-modeling-fundamentals.md
├── 03-diffusion-models-3d-generation.md
├── 04-neural-differentiable-rendering.md
├── 05-text-to-3d-multimodal-conditioning.md
├── 06-neural-fields-nerf-gaussian-splatting.md
├── 07-mesh-generation-topology.md
├── 08-texture-material-appearance.md
├── 09-3d-data-pipelines-datasets.md
├── 10-evaluation-metrics-generative-3d.md
├── 11-training-infrastructure-scaling.md
├── 12-research-methodology-experimentation.md
└── 13-scenario-behavioral.md
- 🏷️ Add "Paper tags" (which questions map to specific influential papers — DreamFusion, Zero-1-to-3, 3D Gaussian Splatting, Point-E/Shap-E, Instant-NGP, etc.)
- 🎬 Add a
/mock-interviewsfolder with full simulated research-presentation and whiteboard-derivation sessions - 🎖️ Add difficulty badges per question
- 📝 Add worked derivations for the diffusion ELBO/score-matching objective and volumetric rendering equation
- 📖 Add a companion cheat-sheet repo comparing 3D representation tradeoffs and generative model family tradeoffs
- 💡 Community-submitted "how I answered this in a real interview" notes
This is meant to be a living, community-curated resource. See CONTRIBUTING.md for the format to follow when submitting a question.
Content is released under the MIT License — free to use, fork, and adapt.
Maintained by @ishandutta2007. Part of a series of "Awesome" curated technical resources.