Research Scientist - Synthetic Data for CAD Foundation Models (Spectral Labs)

failed58 qualified2 runsMar 9, 1:16 PMsynthetic-data-3d-cad-research

Qualified Candidates (58)

Artem Lukoianov

high hireability

Research Intern@DeepMind

Previously: Research Intern @ Meta

Boston, US

SolidGen — only candidate with direct B-Rep solid model generation expertise (ICLR 2024); differentiable parametric CAD work directly maps to Spectral Labs' CAD foundation model mission (h=5)
Hireability: HIGH — MIT CSAIL PhD candidate graduating 2026, DeepMind intern

SolidGen: An Autoregressive Model for Direct B-rep Synthesis MeshSDF: Differentiable Iso-Surface Extraction Score Distillation via Reparametrized DDIM ottogin/SolidGen

Amin Heyrani Nobari

high hireability

Research Assistant@MIT

Previously: Research Assistant @ University of Toronto

Boston, US

MIT DECODE Lab — most directly relevant candidate for engineering design + parametric CAD AI
BikeBench + OAT + NITO trifecta; two NeurIPS 2025 papers
Website rebuilt Dec 2025 = job market entry signal
Hireability: HIGH — prime graduation window

BikeBench: A Bicycle Design Benchmark for Generative Models with Objectives and Constraints NITO: Neural Implicit Fields for Resolution-free Topology Optimization Optimize Any Topology: A Foundation Model for Shape- and Resolution-Free Structural Topology Optimization ahnobari/NITO_Public ahnobari/OptimizeAnyTopology

Lyle Regenwetter

high hireability

Graduate Student@Massachusetts Institute of Technology

Previously: ML Intern, Platform Architecture @ Apple

Boston, US

MIT DECODE Lab — parametric bicycle CAD datasets (BIKED, BIKED++) and deep generative design models (h=8)
CV update Jan 2026 is direct job-market signal
Apple ML intern background
Hireability: HIGH — prime graduation window with explicit job market signal

BIKED: A Dataset for Computational Bicycle Design with Machine Learning Benchmarks Biked++: A Multimodal Dataset of 1.4 Million Bicycle Image and Parametric CAD Designs Deep Generative Models in Engineering Design: A Review Lyleregenwetter/BikeBench

Di Chang

high hireability

Student Researcher@Meta

Previously: Research Scientist Intern @ Meta

San Francisco, US

USC PhD (Soleymani/IHP-Lab) interning at Meta SuperIntelligence Labs in Palo Alto
MagicPose, MagicPose4D dynamic 3D generation (h=8)
Location move to Bay Area in Jan 2026 is explicit post-PhD placement signal
Hireability: HIGH — final-year PhD, Bay Area relocation, Meta MSL internship active

MagicPose: Realistic Human Poses and Facial Expressions Retargeting with Identity-aware Diffusion MagicPose4D: Crafting Articulated Models with Appearance and Motion Control

Yang You

high hireability

Postdoctoral Researcher@Stanford University

Previously: PhD @ Shanghai Jiao Tong University

Stanford, US

Img2CAD (SIGGRAPH Asia 2025) — reverse engineering 3D CAD models from images via VLM-assisted factorization; KeypointNet (CVPR 2020) 3D keypoint understanding; Stanford Guibas Lab postdoc (h=10)
Hireability: HIGH — postdoc at Stanford, standard industry transition window

Img2CAD: Reverse Engineering 3D CAD Models from Images through VLM-Assisted Conditional Factorization qq456cvb/Img2CAD

Aditya Ganeshan

high hireability

PhD Student@Brown University

ParSEL (SIGGRAPH Asia 2024) parametric 3D shape editing + program synthesis for 3D shapes at Brown (h=5)
Neurosymbolic and parametric shape generation overlaps with CAD sequence modeling
NOTE: original Oxford/TextCAD attribution was incorrect — he is at Brown
Hireability: HIGH — Brown PhD nearing completion with strong recent venue placements

ParSEL: Parameterized Shape Editing with Language BardOfCodes

Chuhang Zou

high hireability

Research Scientist@Meta

Previously: Senior Research Scientist @ Amazon

Seattle, US

3D-PRNN shape primitives (277 citations), LayoutNet 3D room layout (327 citations), Direct 3D Generation (2025) at Meta (h=15); website shows 16 recent changes including position updates in 2025
Hireability: HIGH — multiple 2025 position updates and frequent website changes are strong transition signals

3D-PRNN: Generating Shape Primitives with Recurrent Neural Networks LayoutNet: Reconstructing the 3D Room Layout from a Single RGB Image Direct and Explicit 3D Generation from a Single Image

Ghadi Nehme

high hireability

PhD Student@MIT (MechE, DeCoDe Lab)

ex-Stanford University

VideoCAD (NeurIPS 2025) — 41K synthetic CAD UI interaction dataset + VideoCADFormer model at MIT DeCoDe Lab (h=5)
Direct overlap with synthetic CAD data creation and CAD sequence modeling
Hireability: HIGH — MIT PhD student publishing prolifically, prime transition window

VideoCAD: A Dataset and Model for Learning Long-Horizon 3D CAD UI Interactions from Video ghadinehme/VideoCAD

Aliasghar Khani

high hireability

Researcher@Autodesk

Previously: AI Research Internship @ Autodesk

Vancouver, CA

Co-author on WaLa (2024) billion-parameter 3D generative model at Autodesk AI Lab (h=6)
NOTE: original SkexGen attribution is incorrect — SkexGen is by a different Xiang Xu
Hireability: HIGH — early-career Autodesk researcher, high mobility window

Wavelet Latent Diffusion (WaLa): Billion-Parameter 3D Generative Model with Compact Wavelet Encodings AutodeskAILab/WaLa

Congyue Deng

high hireability

Student Researcher@DeepMind

Previously: Research Scientist @ Waymo

San Francisco, US

Vector Neurons (ICCV 2021, 430 citations), NAP (NeurIPS 2023 — articulated 3D object prior) at MIT postdoc; PhD Stanford (Guibas lab); prior Waymo RS and Google Brain (h=12)
Hireability: HIGH — early postdoc, prime transition window

Vector Neurons: A General Framework for SO(3)-Equivariant Networks NAP: Neural 3D Articulated Object Prior

Dejia Xu

high hireability

Research Scientist@Luma AI

Previously: Research Assistant @ University of Texas at Austin

San Francisco, US

Lead author on AGG (3D Gaussian generation), CamCo (3D-consistent video), and DreamCraft3D; RS at Luma AI building 3D/video generation (h=10)
Hireability: HIGH — fresh PhD graduate in first industry role at a startup, natural transition window

AGG: Amortized Generative 3D Gaussians for Single Image to 3D Diffusion4D: Fast Spatial-temporal Consistent 4D generation via Video Diffusion Models aimed-ufa/DreamCraft3D

Anna C. Doris

high hireability

PhD Student@MIT (DeCoDe Lab)

CAD-Coder (IDETC 2025) VLM for CAD code generation + 163K GenCAD-Code dataset at MIT DeCoDe Lab (h=4)
Synthetic CAD data pipeline directly relevant
Hireability: HIGH — MIT PhD student at prime transition window with strong recent publications

CAD-Coder: An Open-Source Vision-Language Model for Computer-Aided Design Code Generation anniedoris/CAD-Coder

Haitao Yang

high hireability

Ph.D. student@University of Texas at Austin

Previously: Research intern @ Waymo

Austin, US

Author of BrepGen (NeurIPS 2023) and SolidGen (ICLR 2024) — the two most-cited direct B-rep CAD generative model papers, co-authored with Autodesk Research (h=10)
UT Austin PhD with Waymo + Amazon internships
Note: original entry incorrectly listed Autodesk as present company — he is a PhD student
Hireability: HIGH — PhD final year, prime transition window

BrepGen: A B-rep Generative Diffusion Model with Structured Latent Geometry SolidGen: An Autoregressive Model for Direct B-rep Synthesis

Honglin Chen

high hireability

PhD student@Columbia University

Previously: Research Intern @ Adobe

Columbia PhD with Adobe + Meta internships — neural implicit geometry and simulation focus (h=8)
Softer CAD fit than MIT DECODE candidates but strong 3D geometry foundations
Hireability: HIGH — ~4yr Columbia PhD nearing completion with industry internship track record

Implicit Neural Spatial Representations for Time-dependent PDEs The Shape Matching Element Method: Direct Animation of Curved Surface Models

Hong-Xing Yu

high hireability

PhD Candidate@Stanford University

Previously: Intern @ Google

WonderWorld (CVPR 2025 Highlight), PhysDreamer (ECCV 2024), ZeroNVS — generative 3D/world modeling at Stanford (h=15)
Hireability: HIGH — Stanford PhD Candidate nearing completion, Google intern history

WonderWorld: Interactive 3D Scene Generation from a Single Image PhysDreamer: Physics-Based Interaction with 3D Objects via Video Generation ZeroNVS: Zero-shot 360-degree View Synthesis from a Single Real Image KovenYu/WonderWorld

Ke Niu

high hireability

Master's student@Phd of Computer Science Fudan University

Previously: Graphic Designer @ Boston University

Los Angeles, US

CReFT-CAD (NeurIPS 2025) RL fine-tuning + TriView2CAD 200K synthetic benchmark, PHT-CAD (2025) VLM for engineering drawing analysis at Fudan University (h=6)
NOTE: original MIT affiliation was incorrect — he is at Fudan
Hireability: HIGH — early-career student, prime window for first industry role

CReFT-CAD: Boosting Orthographic Projection Reasoning for CAD via Reinforcement Fine-Tuning PHT-CAD: Efficient CAD Parametric Primitive Analysis with Progressive Hierarchical Tuning

Kwan-Yee Lin

high hireability

Postdoc@University of Michigan

Previously: Researcher @ Shanghai Artificial Intelligence Laboratory

Ann Arbor, US

SynBody synthetic 3D human dataset + MonoHuman animatable neural human field; postdoc at U Michigan (h=15)
Website actively updated
Hireability: HIGH — postdoc at prime transition window, actively publishing

MonoHuman: Animatable Human Neural Field from Monocular Video Stylegan-Human: A Data-Centric Odyssey of Human Generation

Ningna Wang

high hireability

Postdoctoral Research Scientist@Columbia University

ex-University of Texas at Dallas

CADDreamer (CVPR 2025) B-rep generation + SIGGRAPH-level 3D geometry work at Columbia postdoc (h=8)
WiGRAPH Rising Star 2025 and MIT EECS Rising Stars 2025
Hireability: HIGH — postdoc at prime transition window with top recognition

CADDreamer: CAD Object Generation from Single-view Images MATTopo: Topology-preserving Medial Axis Transform with Restricted Power Diagram ningnawang/MATTopo

R. Kenny Jones

high hireability

Postdoctoral Scholar@Stanford University

ex-Brown University

ShapeAssembly (SIGGRAPH Asia 2020) + ShapeCoder (SIGGRAPH 2023) program synthesis for 3D shapes; Stanford postdoc with Agrawala + Wu (h=8)
Procedural/symbolic 3D shape generation maps directly onto CAD construction sequence modeling
Hireability: HIGH — Stanford postdoc at prime 1-3yr transition window

ShapeAssembly: Learning to Generate Programs for 3D Shape Structure Synthesis ShapeCoder: Discovering Abstractions for Visual Programs from Unstructured Primitives rkjones4/ShapeAssembly rkjones4/ShapeCoder

Ruoshi Liu

high hireability

Research Scientist@Amazon

Previously: Research Scientist @ Meta

San Francisco, US

Zero-1-to-3 and Objaverse-XL co-creator — built the canonical synthetic 3D object dataset and zero-shot 3D generation pipeline (h=12)
RS at Amazon FAR SF
Hireability: HIGH — Meta→Amazon Jan 2026 move confirmed; UMD faculty Sep 2026 creates a defined and urgent industry bridge window

Zero-1-to-3: Zero-Shot One Image to 3D Object Objaverse-XL: A Universe of 10M+ 3D Objects cvlab-columbia/zero123

Yang Fu

high hireability

Research Intern@Waymo

Previously: Research Intern @ NVIDIA

San Francisco, US

COLMAP-Free 3DGS (CVPR 2024 Highlight), HOIDiffusion (CVPR 2024), RGBD Objects in the Wild (CVPR 2024) at UCSD PhD / Waymo intern (h=10)
NVIDIA internships x2
Hireability: HIGH — PhD nearing completion, active internship cycle ongoing

COLMAP-Free 3D Gaussian Splatting HOIDiffusion: Generating Realistic 3D Hand-Object Interaction Data RGBD Objects in the Wild: Scaling Real-World 3D Object Learning from RGB-D Videos

Yifei Li

high hireability

CS PhD Student@Massachusetts Institute of Technology

Previously: Research Scientist Intern @ NVIDIA

Cambridge, US

JoinABLe (CVPR 2022) parametric CAD joint assembly learning, LLMs for engineering design (134 citations), differentiable simulation at MIT CSAIL (h=8)
LinkedIn headline explicitly updated to 'Final-Year CS PhD' in Feb 2026
Hireability: HIGH — explicit final-year PhD signal, industry internships at NVIDIA and The AI Institute

JoinABLe: Learning Bottom-up Assembly of Parametric CAD Joints How Can Large Language Models Help Humans in Design and Manufacturing?

Zhenjia Xu

high hireability

Member of Technical Staff@Genesis AI

Previously: Senior Research Scientist @ NVIDIA

San Francisco, US

Diffusion Policy (2052 citations) + Genesis AI physics engine co-creator — sim-to-real synthetic 3D data pipeline expertise (h=10)
Now MTS at Genesis AI SF after leaving NVIDIA GEAR
Hireability: HIGH — recent move to early-stage startup signals active career transition window

Diffusion Policy: Visuomotor Policy Learning via Action Diffusion Universal Manipulation Interface: In-the-Wild Robot Teaching Without In-the-Wild Robots Genesis-Embodied-AI/Genesis

Zhiyang Dou

high hireability

PhD student@MIT

Previously: Visiting Student @ University of Pennsylvania

Boston, US

Wonder3D single-image-to-3D (614 citations, CVPR 2024), Surf-D diffusion surface generation, Part123 part-aware reconstruction at MIT (h=10)
Hireability: HIGH — active MIT PhD student, 614-citation lead paper signals strong output and near-graduation readiness

Wonder3D: Single Image to 3D Using Cross-Domain Diffusion Surf-D: Generating High-Quality Surfaces of Arbitrary Topologies Using Diffusion Models Part123: Part-aware 3D Reconstruction from a Single-view Image DSaurus/Gaussian4D

Ziyu Guo

high hireability

Research Intern@Meta

Previously: Research Intern @ Roblox

San Francisco, US

PointCLIP (CVPR 2022), Point-M2AE (NeurIPS 2022), HOIDiffusion (CVPR 2024 — synthetic 3D interaction data), COLMAP-Free 3DGS (CVPR 2024) at CUHK PhD / Meta intern (h=8)
Website 67 updates, most recent 13 days ago
Hireability: HIGH — PhD nearing completion, highly active market signal

HOIDiffusion: Generating Realistic 3D Hand-Object Interaction Data Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-training PointCLIP: Point Cloud Understanding by CLIP ZiyuGuo99

Ari Seff

medium hireability

Research Scientist@OpenAI

Previously: Research Scientist @ Apple

SketchGraphs (ICML 2020, 100 citations) + Vitruvion (ICLR 2022, 95 citations) — the two most-cited parametric 2D CAD sketch generative modeling papers
Now at OpenAI (joined Jan 2025, h=12)
Hireability: MEDIUM — highly relevant technically but just joined OpenAI; revisit in 12+ months

SketchGraphs: A Large-Scale Dataset for Modeling Relational Geometry in Computer-Aided Design Vitruvion: A Generative Model of Parametric CAD Sketches PrincetonLIPS/SketchGraphs

Carlos Esteves

medium hireability

Research Scientist@Google

Previously: PhD Candidate @ University of Pennsylvania

New York, US

SO(3) spherical CNNs (653 citations), light field neural rendering, single-mesh diffusion at Google (h=18)
Hireability: MEDIUM — senior long-tenure Google RS, no departure signals

Learning SO(3) Equivariant Representations with Spherical CNNs Light Field Neural Rendering Single Mesh Diffusion Models with Field Latents for Texture Generation

Chaoyang Wang

medium hireability

Research Scientist@Snap

Previously: Graduate Research Assistant @ Carnegie Mellon University

Los Angeles, US

Autodecoding Latent 3D Diffusion, 4Real 4D scene generation, Atom text-to-mesh, VD3D 3D-controllable video at Snap (h=15)
Hireability: MEDIUM — early post-PhD RS at Snap, actively publishing, no strong departure signals yet

Autodecoding Latent 3D Diffusion Models 4Real: Towards Photorealistic 4D Scene Generation via Video Diffusion Models VD3D: Taming Large Video Diffusion Transformers for 3D Camera Control

Chen-Hsuan Lin

medium hireability

Senior Research Scientist@NVIDIA

Previously: Research Scientist @ NVIDIA

San Francisco, US

Magic3D (CVPR 2023), Neuralangelo (CVPR 2023), GET3D (NeurIPS 2022), Cosmos world foundation model — core 3D generative research at NVIDIA (h=20)
Hireability: MEDIUM — entrenched NVIDIA tenure, no departure signal

Neuralangelo: High-Fidelity Neural Surface Reconstruction Cosmos World Foundation Model Platform for Physical AI NVlabs/neuralangelo nv-tlabs/GET3D

Edith Tretschk

medium hireability

Research Scientist@Meta

Previously: PhD Candidate @ Max Planck Institute for Informatics

San Francisco, US

Non-Rigid NeRF (633 citations), PatchNets implicit 3D, SceNeRFlow dynamic scene reconstruction at Meta (h=10)
Hireability: MEDIUM — stable Meta RS post-MPI PhD, no active signals, moderate transition window

Non-Rigid Neural Radiance Fields: Reconstruction and Novel View Synthesis of a Dynamic Scene From Monocular Video PatchNets: Patch-Based Generalizable Deep Implicit 3D Shape Representations SceNeRFlow: Time-Consistent Reconstruction of General Dynamic Scenes

Evan Casey

medium hireability

Research Engineer@Autodesk Research

Fusion 360 Gallery dataset (ACM TOG 2021) + Aligning Constraint Generation (ICCV 2025, first author) at Autodesk Research (h=4)
Directly built the most-used real-world CAD sequence dataset and leads constraint alignment research
Hireability: MEDIUM — stable at Autodesk, but ICCV first-author signals research ambition and transition openness

Fusion 360 Gallery: A Dataset and Environment for Programmatic CAD Construction from Human Design Sequences Aligning Constraint Generation with Design Intent in Parametric CAD AutodeskAILab/Fusion360GalleryDataset

Giljoo Nam

medium hireability

Research Scientist@Meta

Previously: Research Intern @ Meta

Pittsburgh, US

Relightable Gaussian Codec Avatars (2024), CT2Hair (SIGGRAPH 2023), Neural Strands (ECCV 2022) — photorealistic 3D generative assets at Meta Reality Labs (h=12)
Hireability: MEDIUM — stable Meta role, transition window open

Relightable Gaussian Codec Avatars CT2Hair: High-Fidelity 3D Hair Modeling using Computed Tomography Neural Strands: Learning Hair Geometry and Appearance from Multi-view Images

Hao Tan

medium hireability

Research Scientist@Adobe

Previously: Research Intern @ Bloomberg

San Francisco, US

LRM, GS-LRM, DMV3D, MeshLRM — end-to-end 3D asset synthesis from sparse views (h=22)
Adobe RS
Hireability: MEDIUM — ~4 years at Adobe, transition window open but no active signal

LRM: Large Reconstruction Model for Single Image to 3D GS-LRM: Large Reconstruction Model for 3D Gaussian Splatting DMV3D: Denoising Multi-view Diffusion Using 3D Large Reconstruction Model

Hung-Yu Tseng

medium hireability

Research Scientist@Meta

Previously: Graduate Research Assistant @ University of California, Merced

San Francisco, US

Robust Dynamic Radiance Fields, pose-guided diffusion for consistent 3D view synthesis, 3D scene stylization at Meta (h=12)
Hireability: MEDIUM — stable Meta RS, transition window present but no active signals

Robust Dynamic Radiance Fields Consistent View Synthesis with Pose-Guided Diffusion Models

Ka-Hei Hui

medium hireability

Researcher@Autodesk

Previously: PhD student @ The Chinese University of Hong Kong

CUHK PhD → Autodesk AI Lab
Make-A-Shape (10M-scale 3D diffusion, most directly relevant in batch), Neural Wavelet Diffusion, SP-GAN (h=8)
Note: DB listed as CUHK PhD student; web confirms Autodesk AI Lab
Hireability: MEDIUM — working on exact 3D foundation model domain at Autodesk; worth targeted outreach

Make-A-Shape: a Ten-Million-scale 3D Shape Model Neural Wavelet-domain Diffusion for 3D Shape Generation, Inversion, and Manipulation SP-GAN: Sphere-Guided 3D Shape Generation and Manipulation

Lucy Chai

medium hireability

Researcher@Google

Previously: PhD student @ Massachusetts Institute of Technology

San Francisco, US

Persistent Nature (CVPR 2023 — generative unbounded 3D worlds), Generative Image Dynamics (CVPR 2024), MegaSaM (2025), Stereo Magnification (SIGGRAPH 2018, 1316 citations) at Google (h=12)
PhD MIT (Isola lab)
Hireability: MEDIUM — recently joined Google post-PhD, transition window opening

Persistent Nature: A Generative Model of Unbounded 3D Worlds Generative Image Dynamics Stereo Magnification: Learning View Synthesis Using Multiplane Images mega-sam/mega-sam

Matheus Gadelha

medium hireability

Research Scientist@Adobe

Previously: Research Assistant @ University of Massachusetts Amherst

Seattle, US

3D shape generation from sketches/2D views, ANISE neural implicit assembly, GEM3D, DMesh at Adobe Research (h=15)
Hireability: MEDIUM — established Adobe RS with no departure signals, but 3D-native org could attract a move

3d shape reconstruction from sketches via multi-view convolutional networks ANISE: Assembly-Based Neural Implicit Surface Reconstruction GEM3D: GEnerative Medial Abstractions for 3D Shape Synthesis matheusgadelha/MRTNet

Mikaela Angelina Uy

medium hireability

Researcher@NVIDIA

Previously: PhD Student @ Stanford University

San Francisco, US

Point2Cyl (CVPR 2022) and MV2Cyl (NeurIPS 2024) — parametric extrusion cylinder extraction from point clouds and images; co-author Img2CAD (SIGGRAPH Asia 2025)
NVIDIA Researcher from Stanford PhD (Jul 2025, h=10)
Hireability: MEDIUM — recently started at NVIDIA (<1 year, Jul 2025); revisit mid-2026

Point2Cyl: Reverse Engineering 3D Objects from Point Clouds to Extrusion Cylinders MV2Cyl: Reconstructing 3D Extrusion Cylinders from Multi-View Images mikacuy/point2cyl

Mike Roberts

medium hireability

Senior Research Scientist@Adobe

Previously: Research Scientist @ Intel

San Francisco, US

Hypersim (CVPR 2021, 77K synthetic indoor scenes), PISA (ICML 2025, physics post-training via synthetic video) at Adobe (h=15)
Previously at Apple and Intel
Hireability: MEDIUM — stable senior role, but strong synthetic data lineage is directly on-target

Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understanding PISA Experiments: Exploring Physics Post-Training for Video Diffusion Models by Watching Stuff Drop apple/ml-hypersim

Pradeep Kumar Jayaraman

medium hireability

Senior Manager, Research Science@Autodesk Research

UV-Net (CVPR 2021), BRepNet, SolidGen, BrepGen (SIGGRAPH 2024), AutoBrep — comprehensive B-Rep CAD learning portfolio at Autodesk Research (h=15)
Now Senior Manager, Research Science
Hireability: MEDIUM — ~5+ years at Autodesk, management track reduces transition probability, no explicit signal

UV-Net: Learning from Boundary Representations BrepGen: A B-rep Generative Diffusion Model with Structured Latent Geometry SolidGen: An Autoregressive Model for Direct B-rep Synthesis AutodeskAILab/UV-Net AutodeskAILab/BRepNet

Pradyumna Reddy

medium hireability

Senior Research Scientist@Autodesk

Previously: Senior Research Scientist @ Huawei

London, GB

WaLa (2024) billion-parameter 3D generative model, 3D-WAG hierarchical wavelet autoregressive 3D generation, G3DR at Autodesk Research (h=7)
Track record of cross-company mobility (DeepMind→Huawei→Autodesk)
Hireability: MEDIUM — active mover historically, no explicit current signal

Wavelet Latent Diffusion (WaLa): Billion-Parameter 3D Generative Model with Compact Wavelet Encodings 3D-WAG: Hierarchical Wavelet-Guided Autoregressive Generation for High-Fidelity 3D Shapes AutodeskAILab/WaLa

Qiangeng Xu

medium hireability

Research Scientist@Waymo

Previously: Software Engineer @ Google

San Francisco, US

DISN neural implicit 3D reconstruction (649 citations), GaussianFlow 4D Gaussian generation, Point-NeRF at Waymo (h=12)
Hireability: MEDIUM — demonstrated mobility (Google→Waymo), plausible transition window

DISN: Deep Implicit Surface Network for High-quality Single-view 3D Reconstruction GaussianFlow: Splatting Gaussian Dynamics for 4D Content Creation laughtervv/DISN

Ruilong Li

medium hireability

Research Scientist@NVIDIA

Previously: Research Assistant @ Vision & Graphics Lab at USC ICT

nerfacc and gsplat maintainer — foundational open-source 3D rendering libraries (h=15); now RS at NVIDIA Spatial Intelligence Lab
Hireability: MEDIUM — recent company change Jul 2025, under 1yr

Gaussian Masked Autoencoders nerfstudio-project/nerfacc nerfstudio-project/gsplat

Sai Bi

medium hireability

Senior Research Scientist@Adobe

Previously: Research Intern @ Meta

San Francisco, US

LRM, GS-LRM, DMV3D, MeshLRM — defined the large reconstruction model paradigm for 3D asset synthesis at Adobe Research SF (h=28)
Hireability: MEDIUM — stable senior role, approaching natural tenure inflection point

LRM: Large Reconstruction Model for Single Image to 3D GS-LRM: Large Reconstruction Model for 3D Gaussian Splatting DMV3D: Denoising Multi-view Diffusion Using 3D Large Reconstruction Model

Sam Xu

medium hireability

Senior Research Scientist@Autodesk AI Lab

ex-Simon Fraser University

SkexGen (ICML 2022), BrepGen (SIGGRAPH 2024), AutoBrep (SIGGRAPH Asia 2025) — a full lineage of CAD construction sequence and B-Rep generative models at Autodesk AI Lab
NOTE: original record incorrectly listed Waymo — actual affiliation is Autodesk AI Lab (h≈10)
Hireability: MEDIUM — established at Autodesk, transition window unclear, no explicit open signal

BrepGen: A B-rep Generative Diffusion Model with Structured Latent Geometry SkexGen: Autoregressive Generation of CAD Construction Sequences with Disentangled Codebooks samxuxiang/BrepGen samxuxiang/SkexGen

Sharath Girish

medium hireability

Research Scientist@Snap

Previously: Research Intern @ NVIDIA

College Park, US

EAGLES (3D Gaussian splatting compression, ECCV 2024), QUEEN (quantized free-viewpoint video), SHACIRA (neural implicit compression) at Snap (h=8)
Hireability: MEDIUM — recently started at Snap (~1yr); strong 3D representation background relevant to spatial intelligence

EAGLES: Efficient Accelerated 3D Gaussians with Lightweight EncodingS QUEEN: QUantized Efficient ENcoding of Dynamic Gaussians for Streaming Free-viewpoint Videos Sharath-girish/efficientgaussian

Songyou Peng

medium hireability

Senior Research Scientist@DeepMind

Previously: Senior Researcher/PostDoc @ ETH Zürich

San Francisco, US

Convolutional Occupancy Networks pioneer and DepthSplat (CVPR 2025) author — strong feed-forward 3D reconstruction track record (h=20)
Hireability: MEDIUM — stable Senior RS at DeepMind SF, relatively recent join

Convolutional Occupancy Networks DepthSplat: Connecting Gaussian Splatting and Depth NICE-SLAM: Neural Implicit Scalable Encoding for SLAM autonomousvision/convolutional_occupancy_networks cvg/depthsplat pengsongyou/openscene

Willi Menapace

medium hireability

Senior Research Scientist@Snap Inc.

Previously: Research Scientist @ Snap Inc.

Los Angeles, US

4Real (4D scene generation), VD3D (3D camera control), Snap Video at Snap (h=10)
Promoted to Senior RS with full Snap tenure
Hireability: MEDIUM — entrenched at Snap, no open signals, but 4D/dynamic 3D expertise is directly relevant; worth a cold outreach

4Real: Towards Photorealistic 4D Scene Generation via Video Diffusion Models VD3D: Taming Large Video Diffusion Transformers for 3D Camera Control Snap Video: Scaled Spatiotemporal Transformers for Text-to-Video Synthesis

Xianghao Xu

medium hireability

Software Engineer@Waymo

ex-Brown University

Zone Graphs (CVPR 2021) CAD modeling sequence inference + Part Retrieval Assembly (CVPR 2023) at Brown PhD, now SWE at Waymo (h=6)
Strong CAD AI background but non-research current role
Hireability: MEDIUM — recently joined Waymo post-PhD, SWE not research, transition window open

Inferring CAD Modeling Sequences Using Zone Graphs Unsupervised 3D Shape Reconstruction by Part Retrieval and Assembly xxh43/PartRetrievalAndAssembly

Yuan Li

medium hireability

Research Scientist@Autodesk Research

CADDreamer (CVPR 2025, first author) + CadVLM (ECCV 2024) at Autodesk Research (h=8)
B-rep generation and parametric CAD sketch modeling directly match role requirements
Hireability: MEDIUM — stable at Autodesk, 2-4yr tenure signals transition openness

CADDreamer: CAD Object Generation from Single-view Images CadVLM: Bridging Language and Vision in the Generation of Parametric CAD Sketches

Yulia Gryaditskaya

medium hireability

Senior Research Scientist@Adobe

Previously: Assistant Professor in Artificial Intelligence @ University of Surrey

London, GB

LiftConceptSketches3D (SIGGRAPH Asia 2020), OpenSketch product design sketch dataset, sketch-to-3D pipelines at Adobe Research (h=11)
Covers sketch → 3D shape interface central to parametric CAD
Recent website activity
Hireability: MEDIUM — ~3 years at Adobe, prior tenure-track→industry transition indicates openness, website recently active

Lifting Freehand Concept Sketches into 3D OpenSketch: A Richly-Annotated Dataset of Product Design Sketches ygryadit/LiftConceptSketches3D

Zekun Hao

medium hireability

Research Scientist@NVIDIA

Previously: Research Intern @ NVIDIA

San Francisco, US

PointFlow (885 citations), EdgeRunner autoregressive mesh gen, Meshtron large-scale 3D mesh gen, Edify 3D at NVIDIA (h=12)
Hireability: MEDIUM — NVIDIA RS working on production 3D asset generation, no departure signals, but worth outreach

PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows EdgeRunner: Auto-regressive Auto-encoder for Artistic Mesh Generation Meshtron: High-Fidelity, Artist-Like 3D Mesh Generation at Scale ZekunHao/PointFlow

Beichen Li

low hireability

Member of Technical Staff@OpenAI

Previously: Graduate Research Assistant @ MIT

San Francisco, US

MIT CSAIL PhD (Matusik lab) → OpenAI MTS
CAD-Coder (2025) is highly relevant but now entrenched at OpenAI (h=8)
Best approached as a long-horizon prospect or via network connection
Hireability: LOW — ~1yr at OpenAI, high-retention environment

CAD-Coder: An Open-Source Vision-Language Model for Computer-Aided Design Code Generation Match: Differentiable Material Graphs for Procedural Material Capture

Cyril Picard

low hireability

Member of the Technical Staff (Applied AI Researcher)@P-1 AI

Previously: Research Scientist @ Massachusetts Institute of Technology

Boston, US

EPFL PhD → MIT postdoc/RS → P-1 AI MTS
Strong engineering design AI and synthetic dataset expertise (h=6)
Currently too fresh at P-1 AI to be movable
Hireability: LOW — ~7 months at P-1 AI, recently onboarded

From Concept to Manufacturing: Evaluating Vision-Language Models for Engineering Design DATED: Guidelines for Creating Synthetic Datasets for Engineering Design Applications

Guoxian Song

low hireability

Staff Research Scientist@ByteDance

Previously: Senior Research Scientist @ ByteDance

San Francisco, US

PanoHead (CVPR 2023), AvatarGen (ECCV 2022), GETAvatar — 3D generative human avatars at ByteDance (h=12)
Promoted RS→SRS→Staff internally
Hireability: LOW — long entrenched tenure at ByteDance, no departure signals

Panohead: Geometry-aware 3D Full-Head Synthesis in 360°AvatarGen: a 3D Generative Model for Animatable Human Avatars

Matt Deitke

low hireability

AI Researcher@Meta

Previously: Co-founder @ Vercept

San Francisco, US

Objaverse (800K objects) and Objaverse-XL (10M+ objects) creator — defined synthetic 3D dataset infrastructure for the field (h=22)
Now AI Researcher at Meta SF
Hireability: LOW — joined Meta mid-2025 with large compensation package, under 1 year tenure

Objaverse: A Universe of Annotated 3D Objects Objaverse-XL: A Universe of 10M+ 3D Objects ProcTHOR: Large-Scale Embodied AI Using Procedural Generation allenai/objaverse-xl

Pratul P. Srinivasan

low hireability

Research Scientist@DeepMind

Previously: Intern @ Google

San Francisco, US

NeRF co-inventor (h=39); SimVS and Bolt3D are directly on-topic synthetic 3D data pipelines for training 3D generative models
Hireability: LOW — SIGGRAPH 2025 award winner, entrenched at DeepMind SF

Nerf: Representing Scenes as Neural Radiance Fields for View Synthesis Bolt3D: Generating 3D Scenes in Seconds Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields

Rundi Wu

low hireability

Research Scientist@DeepMind

Previously: PHD Student @ Columbia University

San Francisco, US

DeepCAD — seminal generative transformer for CAD command sequences (ICCV 2021); foundational CAD generation work (h=15)
Hireability: LOW — recently joined DeepMind SF (Aug 2025), under 1 year tenure

DeepCAD: A Deep Generative Network for Computer-Aided Design Models PQ-NET: A Generative Part Seq2Seq Network for 3D Shapes ChrisWu1997/DeepCAD rundiwu/PQ-NET

Runs

#2completed26 qualified / 26 foundMar 12, 8:02 PM

#1completed32 qualified / 32 foundMar 12, 8:02 PM