ML Data Scientist
Are you passionate about making ML models more transparent? Explaining when and how a ML model could fail with tests and statistics? Are you passionate about being a gate keeper for ML models that might impact millions of people? Are you passionate about the latest development in 3D Computer Vision and Graphics?
The Data Analytic and Quality (DAQ) group in Hong Kong is seeking a Machine Learning Data Scientist specializing in 3D generation models to evaluate, optimize, and interpret machine learning models used for 3D content creation.
In this role, you will analyze model performance, identify weaknesses, and improve robustness, efficiency, and realism in generated 3D assets. You will work closely with researchers, engineers, and product teams to ensure high-quality outputs in areas such as 3D reconstruction, generative modeling, and rendering.
If you are passionate about 3D generative AI and pushing the boundaries of AI-driven 3D content creation, we would love to hear from you!
- Analyze and evaluate 3D generation models (e.g., NeRF, GANs, diffusion models, implicit neural representations..etc).
- Develop and apply quality metrics for 3D models quality (e.g., Chamfer Distance, Intersection-over-Union (IoU), FID, PSNR, SSIM).
- Investigate model weaknesses, bias, and generalization using techniques such as SHAP, LIME, and adversarial testing.
- Optimize models for efficiency, realism, and scalability in real-world applications.
- Conduct A/B testing and statistical validation of improvements in 3D model generation.
- Work with 3D data (point clouds, meshes, voxels, signed distance functions) and preprocessing techniques.
- Collaborate with computer vision researchers, graphics, ML engineers, to enhance 3D generation models.
- Utilize MLOps and model tracking tools (e.g., MLflow, Weights & Biases) to monitor model performance over time.
- Bachelor's/Master's/PhD degree in a relevant field such as Computer Science, Data Science, Artificial Intelligence, or related disciplines.
- Open to both recent graduates and professionals with industry experience.
- Solid mathematical background in linear algebra, optimization, and statistical analysis.
- Proficiency in Python and ML libraries (NumPy, Pandas, Matplotlib, Seaborn).
- Experience with ML frameworks and tools: TensorFlow, PyTorch, Scikit-Learn,
- Experience with cloud platforms (AWS, GCP, Azure) for training and deploying large-scale model.
- Knowledge of generative models (GANs, VAEs, Score-Based Models) and 3D reconstruction techniques.
- Excellent verbal and written communication abilities to collaborate across technical and non-technical teams.
- Strong understanding of 3D deep learning techniques, including NeRF, PointNet, MeshCNN, and Diffusion Models.
- Hands-on experience with 3D data processing and rendering frameworks (e.g., Blender API, Unreal Engine, Three.js, Open3D, Kaolin, Minkowski Engine)
- Understanding of feed-forward reconstruction networks for 3D tasks
- Experience with real-time rendering and physics-based simulations.
- Knowledge of edge AI and efficient 3D model compression techniques
- Proven ability to design and optimize complex ML pipelines or AI-driven solutions.