RCI-ABBV-34257 Computer Vision Engineer (3D Imaging/3D Digital Endpoint/3D Data Format/Machine Learning/Photogrammetry/PyTorch/Python/NumPy/SciPy/LLM/CNN/Yolo)

3D Data Scientist

Location/Site: Irvine, CA - Hybrid

Role Summary

We are seeking a highly skilled and intellectually curious 3D Data Scientist to join our growing Digital Endpoints team at the intersection of computational science, facial aesthetics, and cutting-edge 3D capture technology.

• This is a pioneering role: you will be among the first members of the team to operationalize 3D data science capabilities, building on a strong foundation of over 100 validated digital endpoints already developed for 2D images and video.
• In this role, you will lead the validation of a state-of-the-art 3D capture system, architect robust validation pipelines using photogrammetry-rendered 3D imagery, and collaborate cross-functionally to define and develop the next generation of 3D digital endpoints in the facial region for aesthetics applications.
• You will sit at the convergence of machine learning, 3D rendering, and scientific rigor — and your work will directly shape how aesthetic outcomes are measured, quantified, and communicated in clinical and commercial settings.
• This role is equal parts scientist and builder. You must move fluidly between data science workflows and 3D rendering environments, think with both precision and product-mindedness, and bring a strong bias toward innovation without sacrificing scientific integrity.

Key Responsibilities

3D Capture Validation

• Lead the end-to-end validation of a 3D facial capture system, establishing technical benchmarks for accuracy, repeatability, and clinical relevance.
• Design and execute structured validation pipelines using 3D rendered photogrammetry images to evaluate system capabilities across diverse subject populations and capture conditions.
• Develop quantitative test protocols and statistical frameworks to assess 3D capture fidelity, geometric precision, and landmark reproducibility.
• Document findings with scientific rigor and communicate validation outcomes to technical and non-technical stakeholders.

3D Digital Endpoint Development

• Partner with the Digital Endpoints team to define, prototype, and scale a new suite of 3D digital endpoints for facial aesthetics applications, extending the team's existing library of 100+ 2D endpoints.
• Translate 3D capture capabilities and mesh data into clinically meaningful, computable biomarkers and outcome measures.
• Drive hypothesis generation and experimental design for novel 3D endpoints, balancing scientific validity with practical scalability.
• Establish best practices for 3D data preprocessing, surface reconstruction quality control, and feature extraction pipelines.

Machine Learning & Modeling

• Build and evaluate machine learning models (supervised, self-supervised, and geometric deep learning) applied to 3D facial meshes, point clouds, and photogrammetry assets.
• Design experiments to benchmark model performance, generalizability, and robustness across capture systems and patient demographics.
• Iterate rapidly on model architecture and training strategies in close collaboration with engineering and science teams.

Cross-Functional Collaboration & Innovation

• Serve as the technical bridge between the data science team and 3D rendering/capture specialists, translating requirements bidirectionally with clarity and precision.
• Collaborate with clinical scientists, product managers, and regulatory stakeholders to ensure endpoints are fit-for-purpose in aesthetic clinical trials and commercial applications.
• Champion a culture of experimentation, reproducibility, and continuous improvement across 3D data science workflows.
• Stay ahead of the curve on emerging tools, techniques, and literature in 3D computer vision, neural rendering, and digital biomarkers.

Required Qualifications

• Bachelor's degree or higher in Computer Science, Data Science, Computational Biology, Biomedical Engineering, Computer Vision, or a closely related quantitative field (Master's or PhD strongly preferred).
• 3+ years of hands-on experience in data science or machine learning roles, with a demonstrated track record of delivering production-quality work.
• Strong proficiency in Python and standard data science libraries (NumPy, SciPy, Pandas, scikit-learn, PyTorch or TensorFlow).
• Demonstrable experience working with 3D data formats — including meshes, point clouds, depth maps, or photogrammetry outputs — in a research or applied context.
• Deep familiarity with at least one professional 3D rendering or modeling platform such as Blender, Autodesk Maya, or equivalent.
• Proven ability to design and execute rigorous validation or benchmarking studies with a statistical foundation.
• Strong written and verbal communication skills, with the ability to present complex technical findings to diverse audiences.
• Comfortable operating in ambiguous, fast-moving environments with a high degree of autonomy and ownership.

Preferred Qualifications

• Experience in the aesthetics, dermatology, medical imaging, or clinical digital health domain.
• Familiarity with photogrammetry pipelines and tools (e.g., RealityCapture, Agisoft Metashape, or similar).
• Exposure to geometric deep learning frameworks (e.g., PyTorch Geometric, Open3D, trimesh).
• Experience developing digital endpoints, biomarkers, or outcome measures in a regulated or clinical context.
• Knowledge of 3D facial landmarking, surface parameterization, or shape analysis methods.
• Experience contributing to or leading cross-functional research and development projects in an industry setting.
• Familiarity with version control, MLOps principles, and reproducible research practices (Git, DVC, MLflow, or equivalent).

Technical Skills

Programming

• Python (primary), R (secondary), SQL

Machine Learning

• PyTorch, TensorFlow, scikit-learn, geometric deep learning

3D Rendering & Modeling

• Blender, Autodesk Maya, RealityCapture, Metashape

3D Data Processing

• Open3D, trimesh, PyMeshLab, PCL, point cloud / mesh workflows

Photogrammetry

• 3D mesh reconstruction, UV mapping, texture baking, depth maps

Validation & Statistics

• A/B testing, ICC, Bland-Altman analysis, bootstrap methods

Data Infrastructure

• Git, DVC, MLflow, cloud platforms (AWS / GCP / Azure)

Visualization

• Matplotlib, Plotly, ParaView, 3D scene rendering pipelines

Success Metrics / What Success Looks Like in the First 12 Months

We believe in setting clear expectations. Here is what exceptional performance looks like in this role :

0–3 Months

• Deep understanding of the existing 2D digital endpoint library, data infrastructure, and 3D capture system architecture. Initial validation protocol drafted and under review.

3–6 Months

• First full validation pipeline for the 3D capture system is operational. Preliminary validation report delivered to scientific and product stakeholders with clear capability characterization.

6–9 Months

• At least 3–5 novel 3D digital endpoint candidates identified, defined, and in active development or prototyping. Machine learning models benchmarked against defined performance criteria.

9–12 Months

• First set of 3D digital endpoints validated, documented, and ready for deployment in an aesthetics study or product context. Cross-functional best practices for 3D data science workflows established and socialized across the team.