I am currently a Research Resident at Qualcomm AI Research under the supervision of Prof. Binh-Son Hua. I received my Bachelor’s degree in Computer Science (APCS) from the University of Science, Vietnam National University Ho Chi Minh City.

My research interests focus on computer vision, computer graphics, and generative models. I am interested in simulating realistic 3D worlds using physically grounded generative models that integrate geometry, appearance, and motion.

News

• 2025.02: Transitioned to Qualcomm AI Research as a Research Resident, focusing on open-vocabulary 3D understanding and physically grounded generative models.
• 2023.10: Won the Vietnam Female Students in Science and Technology Award (Top 20 nationwide).
• 2023.08: Joined VinAI Research as a Research Resident, working on 3D/4D modeling and neural scene representations.

Experience

• 2025.02 - Present, Research Resident at Qualcomm AI Research
• 2025.08 - 2025.10, AI Engineer at CloudThinker.
• 2023.08 - 2025.02, Research Resident at VinAI Research.
• 2021.06 - 2021.12, Data Scientist Intern at AI Lab, MTI Technology.

Publications

CVPR 2026 (Under Review)

Prepare Lighter and Faster for Open-Vocabulary Queries: A Query-Wise 3D Segmenter for Gaussian Splatting

Nhat-Quynh Le-Pham, Khoi Nguyen, Binh-Son Hua

• Proposed a novel on-device, lightning-fast query-wise open-vocabulary 3D instance segmentation framework for 2D/3D Gaussian Splatting.
• Achieved state-of-the-art runtime efficiency and visual quality.

Bachelor Thesis 2024

Text-to-4D Content Creation using Image Priors and Monocular Driver Videos

Nhat-Quynh Le-Pham

Thesis Video

• Developed an end-to-end framework for controllable, high-quality 4D object generation from text prompts using only image priors and monocular driver videos, removing reliance on heavy video diffusion models.
• Successfully defended with a perfect score (10/10).

Patents

• Query-based Open-vocabulary 3D object segmentation from Gaussian Splats, Nhat-Quynh Le-Pham, Khoi Nguyen, Binh-Son Hua, US Patent

Workshops & Journals

3DOR 2023

SketchANIMAR: Sketch-Based 3D Animal Fine-Grained Retrieval

Published in Computers & Graphics (Elsevier, Q1)

Paper Code

• Developed a cross-domain contrastive learning framework aligning 3D objects with sketches and text via cosine similarity using an EfficientNetV2-Small-based feature extraction pipeline with Canny edge + KMeans-based dataset augmentation.

3DOR 2023

TextANIMAR: Text-Based 3D Animal Fine-Grained Retrieval

Published in Computers & Graphics (Elsevier, Q1)

Paper Code

• Built a contrastive learning framework that aligns 3D objects and text using CLIP and EfficientNetV2-Small features, with KMeans dataset augmentation, improving Nearest Neighbor retrieval performance.

3DOR 2022

Fitting and Recognition of Simple Geometric Primitives on Point Clouds

Published in Computers & Graphics (Elsevier, Q1)

Paper Code

• Proposed a PointNet-based approach for geometric primitive recognition in point clouds, employing least squares fitting and majority voting for robust classification.

Honors and Awards

• 2023.10 Awarded the Vietnam Female Students in Science and Technology Award (Top 20 nationwide).
• 2023.03 Received the Top GPA Scholarship from the Faculty of Information Technology, University of Science (Top 1%).
• 2023.03 First Prize, SHREC 2023 — Sketch-Based 3D Animal Fine-Grained Retrieval (3DOR’23).
• 2023.03 First Prize, SHREC 2023 — Text-Based 3D Animal Fine-Grained Retrieval (3DOR’23).
• 2022.02 3rd Place, SHREC 2022 — Fitting and Recognition of Geometric Primitives (3DOR’22).
• 2021.12 Finalist, Ho Chi Minh City AI Challenge.

Professional Services

• Reviewers: CVPRW 2025, ICCV 2025, ICLR 2026.