Back to dashboard

Spectral is building generative foundation models that turn user input into stru…

completed33 qualified1 runMay 4, 8:08 PMspectral-is-building-generative-foundation-models-that-turn-1777925309
ParsedSpectral · 4 topics · Senior · Hybrid
Generating seed nodes
0 proposed
Explored 0 queries
0/0 done
    3
    Expanding nodes
    queued
    4
    Qualifying candidates
    queued

    Qualified Candidates (28)

    FZ

    Fuyang Zhang

    high hireability

    Research Assistant@Simon Fraser University

    Previously: Intern @ Meta

    Burnaby, CA

    72
    3D Geometry Representations82
    B-REP Modeling80
    Geometric Deep Learning75
    Generative CAD Models50
    Strengths
    BR-DF (2025): B-rep volumetric distance functions — direct B-rep CAD work
    SceneScript (ECCV 2024): autoregressive structured 3D scene reconstruction
    Gaps
    B-rep work is representation-focused, not full generative parametric CAD (feature trees, STEP)
    …click to see all
    MF

    Marco Fumero

    high hireability

    Postdoctoral Researcher@Institute of Science and Technology Austria

    Previously: Doctoral Researcher @ Sapienza Università di Roma

    Vienna, AT

    39
    Geometric Deep Learning72
    3D Geometry Representations65
    Generative CAD Models15
    B-REP Modeling5
    Strengths
    ISTA postdoc in Rodolà lab — top European geometric DL group
    CLIP-Forge: zero-shot text-to-3D shape generation (co-author with Sanghi)
    Gaps
    No B-rep, NURBS, or parametric/structured CAD experience
    …click to see all
    RX

    Rui Xu

    high hireability

    Ph.D. Student@The University of Hong Kong

    Previously: Research Intern @ Tencent Visvise

    Hong Kong, HK

    60
    3D Geometry Representations88
    Geometric Deep Learning80
    Generative CAD Models52
    B-REP Modeling20
    Strengths
    NeurCADRecon: CAD surface reconstruction with zero Gaussian curvature constraint
    MeshMosaic (CVPR 2026): generative model for artist-quality mesh assembly
    Gaps
    No explicit structured/parametric CAD generation (B-rep, feature trees, STEP)
    …click to see all
    SX

    Sam Xu

    high hireability
    96
    B-REP Modeling99
    Generative CAD Models99
    3D Geometry Representations95
    Geometric Deep Learning92
    Strengths
    BrepGen (SIGGRAPH 2024): first diffusion model for B-rep with doubly-curved surfaces
    AutoBrep (SIGGRAPH Asia 2025): decoder-only transformer autoregressive B-rep generation
    Gaps
    Autodesk-centric background — unclear exposure to startup pace or production deployment outside Autodesk
    …click to see all
    AS

    Aditya Sanghi

    medium hireability

    Director@Samruddhi

    Previously: Director @ Sanghi Industries Ltd

    Ahmedabad, IN

    96
    Generative CAD Models98
    B-REP Modeling97
    3D Geometry Representations95
    Geometric Deep Learning92
    Strengths
    SolidGen (2023): autoregressive B-rep synthesis — directly query-aligned, 101 citations
    BRepNet (2021): topological message passing on solid models, 177 citations
    Gaps
    No RL-for-generative-models work (RLHF, DPO, GRPO) — JD lists as a plus
    …click to see all
    AR

    Arianna Rampini

    medium hireability

    Senior Research Scientist@Autodesk

    Previously: Junior Research Scientist @ Autodesk

    IT

    50
    Geometric Deep Learning82
    3D Geometry Representations78
    Generative CAD Models28
    B-REP Modeling10
    Strengths
    WaLa (2024): billion-parameter 3D diffusion model, wavelet encodings
    Make-A-Shape: generative model over 10M 3D shapes
    Gaps
    No structured/parametric CAD output — work targets meshes/wavelets, not B-rep or feature trees
    …click to see all
    BM

    Baorui Ma

    medium hireability

    Researcher@Beijing Academy of Artificial Intelligence

    Previously: PhD student @ Tsinghua University

    Beijing, CN

    50
    3D Geometry Representations90
    Geometric Deep Learning85
    Generative CAD Models20
    B-REP Modeling5
    Strengths
    See3D (CVPR 2025 Highlight) — generative 3D creation from video at scale
    UDiff + GeoDream — 3D diffusion generative model experience
    Gaps
    No CAD/B-rep/feature tree experience — not parametric CAD domain
    …click to see all
    CL

    Chao Liu

    medium hireability

    Researcher@NVIDIA

    Previously: Research And Development Scientist @ Elephas

    Madison, US

    35
    3D Geometry Representations68
    Geometric Deep Learning52
    Generative CAD Models15
    B-REP Modeling3
    Strengths
    BlobGEN-3D: compositional 3D-consistent generation with structured blob repr (2024)
    Not-So-Optimal Transport Flows for 3D PCG — generative model for point clouds (2025)
    Gaps
    No CAD/B-rep/parametric output experience — 3D generation is images/point clouds only
    …click to see all
    DL

    Difan Liu

    medium hireability

    Research Scientist@Adobe

    Previously: PhD student @ University of Massachusetts, Amherst

    55
    3D Geometry Representations75
    Geometric Deep Learning65
    Generative CAD Models60
    B-REP Modeling20
    Strengths
    CSGNet — neural CSG parsing, core structured CAD representation (279 citations)
    ParSeNet — NURBS-style parametric surface fitting from 3D point clouds (214 citations)
    Gaps
    Recent 2023–2025 focus: video/image diffusion, not 3D CAD
    …click to see all
    HS

    Hooman Shayani

    medium hireability

    Principal Researcher@Autodesk

    London, GB

    90
    B-REP Modeling95
    3D Geometry Representations92
    Geometric Deep Learning88
    Generative CAD Models85
    Strengths
    BRepNet (CVPR 2021, 159 cit) — topological GNN directly on B-rep solid models
    UV-Net (CVPR 2021, 125 cit) — ML on B-rep boundary curves and surfaces
    Gaps
    Senior Manager role — moving to IC position at startup may be a step back
    …click to see all
    JL

    Joe Lambourne

    medium hireability
    99
    B-REP Modeling100
    Generative CAD Models100
    3D Geometry Representations100
    Geometric Deep Learning95
    Strengths
    BrepGen (SIGGRAPH 2024): diffusion generative model for structured B-rep CAD
    AutoBrep (2025): autoregressive B-rep generation, unified topology+geometry tokenization
    Gaps
    No open-to-work signals; likely entrenched at Autodesk Research
    …click to see all
    KH

    Ka-Hei Hui

    medium hireability

    Researcher@Autodesk

    Previously: PhD student @ The Chinese University of Hong Kong

    US

    47
    3D Geometry Representations80
    Geometric Deep Learning75
    Generative CAD Models25
    B-REP Modeling8
    Strengths
    Make-A-Shape (first author, 2024) — 10M-scale generative 3D shape model
    Neural Wavelet Diffusion (SIGGRAPH Asia 2022, 144 cites) — diffusion over SDF shapes
    Gaps
    No B-rep or parametric/feature-tree CAD publications found
    …click to see all
    KM

    Kamal Rahimi Malekshan

    medium hireability

    Principal Machine Learning Engineer@Autodesk

    Previously: Senior Research Engineer - Machine Learning @ Autodesk

    Toronto, CA

    67
    Generative CAD Models82
    Geometric Deep Learning80
    3D Geometry Representations75
    B-REP Modeling30
    Strengths
    WaLa: billion-param diffusion 3D generative model with wavelet encodings (2024)
    Make-A-Shape: 10M-scale 3D shape generative model (2024)
    Gaps
    No explicit B-rep / NURBS work; editable CAD paper uses voxels, not B-rep
    …click to see all
    KM

    Kanika Madan

    medium hireability

    Principal Researcher@Autodesk

    Previously: Senior Manager Business Solutions @ AU

    Gurugram, IN

    44
    3D Geometry Representations70
    Geometric Deep Learning55
    Generative CAD Models45
    B-REP Modeling5
    Strengths
    WaLa: billion-parameter 3D diffusion model (SDF+wavelet latents, 2024)
    Image-to-3D texture mapping using triplane representations (2025)
    Gaps
    3D work is SDF/implicit-based, not parametric CAD or feature trees
    …click to see all
    KW

    Karl D.D. Willis

    medium hireability
    98
    B-REP Modeling99
    Generative CAD Models99
    3D Geometry Representations98
    Geometric Deep Learning97
    Strengths
    BrepGen (ACM Trans. Graph. 2024) — B-rep generative diffusion model
    SolidGen (TMLR 2022) — autoregressive direct B-rep synthesis
    Gaps
    Recently moved to management (Head of Model Delivery, 2025) — may prefer leadership over IC research
    …click to see all
    ND

    Nishkrit Desai

    medium hireability

    Researcher@Axiom

    Previously: Intern @ NVIDIA

    87
    B-REP Modeling95
    Generative CAD Models95
    3D Geometry Representations85
    Geometric Deep Learning72
    Strengths
    SolidGen: autoregressive B-rep synthesis, ICLR 2024 + TMLR (89 citations)
    Patent: direct B-rep synthesis via ML (US 2024) — exact domain match
    Gaps
    Junior-to-mid seniority (h_index=3, ~3–4 yrs experience)
    …click to see all
    PJ

    Pradeep Kumar Jayaraman

    medium hireability
    94
    B-REP Modeling98
    Generative CAD Models98
    3D Geometry Representations90
    Geometric Deep Learning88
    Strengths
    BrepGen (SIGGRAPH 2024): flagship B-rep diffusion generative model
    AutoBrep (SIGGRAPH Asia 2025): autoregressive B-rep — topology + geometry unified
    Gaps
    No explicit RL-for-generative-models work (RLHF/DPO/GRPO) in publication list
    …click to see all
    SY

    Siming Yan

    medium hireability

    Researcher@Yinwang Intelligent Technology Co. Ltd.

    Previously: PhD student @ The University of Texas at Austin

    41
    3D Geometry Representations78
    Geometric Deep Learning72
    Generative CAD Models10
    B-REP Modeling3
    Strengths
    HPNet (ICCV 2021): primitive segmentation on 3D point clouds
    IAE (ICCV 2023): implicit surfaces for 3D self-supervised learning
    Gaps
    No generative CAD or structured parametric output experience
    …click to see all
    AB

    Adrien Bousseau

    low hireability

    Researcher@INRIA

    Previously: Postdoc @ UC Berkeley

    Nice, FR

    46
    Generative CAD Models65
    3D Geometry Representations55
    Geometric Deep Learning50
    B-REP Modeling15
    Strengths
    Free2CAD (SIGGRAPH 2022) — drawing→CAD commands, Best Paper HM
    Sketch2CAD (2020, 106 cites) — sequential CAD modeling
    Gaps
    No transformer/diffusion generative model experience — not foundation model background
    …click to see all
    AJ

    Alec Jacobson

    low hireability

    Researcher@Adobe

    Previously: Senior Research Scientist @ Adobe

    Toronto, CA

    55
    3D Geometry Representations95
    Geometric Deep Learning85
    B-REP Modeling20
    Generative CAD Models20
    Strengths
    libigl creator — world's most-used geometry processing library
    Canada Research Chair in Geometry Processing at UofT
    Gaps
    No generative CAD or parametric/feature-tree model work
    …click to see all
    AT

    Andrea Tagliasacchi

    low hireability

    Associate Professor (Status Only)@University of Toronto

    Previously: Staff Research Scientist @ DeepMind

    Burnaby, CA

    53
    3D Geometry Representations90
    Geometric Deep Learning82
    Generative CAD Models28
    B-REP Modeling12
    Strengths
    BSP-Net: generative structured mesh output, 381 citations (CVPR 2020)
    Pie-Net: parametric edge inference on point clouds (NeurIPS 2020)
    Gaps
    No direct CAD/B-rep/STEP or feature-tree generative work
    …click to see all
    JL

    Joe Lambourne

    low hireability
    97
    B-REP Modeling99
    Generative CAD Models99
    Geometric Deep Learning95
    3D Geometry Representations95
    Strengths
    AutoBrep (SIGGRAPH Asia 2025): autoregressive transformer for B-rep topology+geometry tokens
    BrepGen (SIGGRAPH 2024): hierarchical diffusion model for B-rep structured latent geometry
    Gaps
    Deeply entrenched at Autodesk (15+ years) — switching risk very high
    …click to see all
    MF

    Matthew Fisher

    low hireability

    Principal Scientist@Adobe

    Previously: Graduate student @ Stanford University

    43
    3D Geometry Representations80
    Geometric Deep Learning72
    Generative CAD Models15
    B-REP Modeling5
    Strengths
    ShapeShifter (CVPR 2025): multiscale point-voxel diffusion for 3D shape variations
    DMesh++: differentiable mesh targeting complex 3D geometry (2025)
    Gaps
    No CAD-specific work: no B-rep, feature trees, parametric or editable outputs
    …click to see all
    PW

    Peng-Shuai Wang

    low hireability

    Assistant Professor@Peking University

    Previously: Researcher @ Microsoft

    Beijing, CN

    56
    3D Geometry Representations95
    Geometric Deep Learning92
    Generative CAD Models30
    B-REP Modeling5
    Strengths
    OctGPT (SIGGRAPH 2025) — autoregressive 3D shape generation, transformer-based
    OctFusion (SGP 2025) — diffusion model for 3D shapes over octree structure
    Gaps
    No B-rep, NURBS, or parametric/structured CAD experience
    …click to see all
    PJ

    Pradeep Kumar Jayaraman

    low hireability
    96
    B-REP Modeling99
    Generative CAD Models98
    3D Geometry Representations95
    Geometric Deep Learning93
    Strengths
    BrepGen (SIGGRAPH 2024): first diffusion model for structured B-rep
    AutoBrep (SIGGRAPH Asia 2025): autoregressive B-rep topology+geometry
    Gaps
    10+ year Autodesk tenure; no open-to-work signals anywhere
    …click to see all
    RW

    Rundi Wu

    low hireability

    Research Scientist@DeepMind

    Previously: PHD Student @ Columbia University

    San Francisco, US

    81
    Generative CAD Models95
    3D Geometry Representations85
    Geometric Deep Learning80
    B-REP Modeling65
    Strengths
    DeepCAD (ICCV 2021): canonical generative parametric CAD sequence model
    PQ-NET (CVPR 2020): generative part Seq2Seq for 3D shapes
    Gaps
    Started Google DeepMind June 2025 — low hireability, <1 year in role
    …click to see all
    XT

    Xin Tong

    low hireability

    Partner Research Manager@Anuttacon

    Previously: Partner Research Manager @ Microsoft

    San Francisco, US

    86
    Geometric Deep Learning95
    3D Geometry Representations95
    B-REP Modeling80
    Generative CAD Models72
    Strengths
    ComplexGen (2022): B-rep chain complex generation from point clouds
    Structured 3D Latents (2025): state-of-art scalable 3D generative model
    Gaps
    No evidence of parametric/feature-tree CAD synthesis (structured editing, not just B-rep reconstruction)
    …click to see all
    YL

    Yang Liu

    low hireability

    Principal Researcher@Microsoft

    Previously: Postdoctoral Fellow @ LORIA/INRIA

    China

    87
    B-REP Modeling95
    3D Geometry Representations92
    Geometric Deep Learning88
    Generative CAD Models72
    Strengths
    ComplexGen (SIGGRAPH 2022) — ML B-rep CAD reconstruction, most cited work in area
    Neural Halfspace B-rep solid conversion (SIGGRAPH Asia 2022)
    Gaps
    Reconstruction-focused (scan→B-rep), not user-input generative CAD
    …click to see all

    Runs

    #1completed0 qualified / 0 foundMay 4, 8:08 PM