Ali Athar

Ali Athar

Applied Scientist

Amazon

Biography

I work as an Applied Scientist at Amazon Frontier AI and Robotics (FAR) and am based out of Berlin, Germany. Previously, I worked as a Research Scientist at ByteDance Seed. I completed my PhD in Computer Science from RWTH Aachen University (Germany) under the supervision of Prof. Bastian Leibe. My research interests include video understanding, spatial intelligence, and object tracking and segmentation.

Interests
  • Computer Vision
  • Video Understanding
  • Object Tracking and Segmentation
Education
  • PhD in Computer Science, 2019 - 2024

    RWTH Aachen University (Aachen, Germany)

  • MSc. in Communications Engineering, 2016 - 2019

    Technical University of Munich (Germany)

  • BE in Electrical Engineering, 2012 - 2016

    National University of Sciences and Technology (Islamabad, Pakistan)

Publications

(2024). Multimodal Four-dimensional Panoptic Segmentation. In US Patent App. 18/736,525.

PDF Cite Project

(2023). 4D-Former: Multimodal 4D Panoptic Segmentation. In CoRL.

PDF Cite Project Video

(2023). TarViS: A Unified Approach for Target-based Video Segmentation. In CVPR (Highlight).

PDF Cite Code Video

(2023). BURST: A Benchmark for Unifying Object Recognition, Segmentation and Tracking in Video. In WACV.

PDF Cite Code Dataset

(2022). HODOR: High-level Object Descriptors for Object Re-segmentation in Video Learned from Static Images. In CVPR (Oral).

PDF Cite Code Video

(2022). Differentiable Soft-Masked Attention. In CVPR-W.

PDF Cite Code

(2022). D^2-Conv3D: Dynamic Dilated Convolutions for Object Segmentation in Videos. In WACV.

PDF Cite Code

(2021). A Single-Stage, Bottom-up Approach for Occluded VIS using Spatio-temporal Embeddings. In ICCV-W.

PDF Cite

(2020). Making a Case for 3D Convolutions for Object Segmentation in Videos. In BMVC.

PDF Cite Code Video

(2020). STEm-Seg: Spatio-temporal Embeddings for Instance Segmentation in Videos. In ECCV.

PDF Cite Code Video

(2019). Whole-body motion and footstep planning for humanoid robots with multi-heuristic search. Robotics and Autonomous Systems.

PDF Cite

(2016). Whole-body motion planning for humanoid robots with heuristic search. In IROS.

PDF Cite DOI

Experience

 
 
 
 
 
Applied Scientist
Amazon FAR
Aug 2025 – Present Berlin, Germany
 
 
 
 
 
Research Scientist
ByteDance Inc.
Feb 2024 – Jul 2025 San Jose, USA
 
 
 
 
 
Research Intern
Jan 2023 – Jun 2023 Toronto, Canada
 
 
 
 
 
Conference Reviewer
Jun 2019 – Present
ICRA 2019, BMVC 2020, CVPR 2021, ICCV 2021, BMVC 2021, CVPR 2022, ECCV 2022, CVIU
 
 
 
 
 
Teaching Assistant
RWTH Aachen
Oct 2019 – Feb 2024
Machine Learning (Winter 2019 and Winter 2020), Computer Vision (Winter 2021)
 
 
 
 
 
Working Student 3D Robotics and Perception
Feb 2017 – Mar 2019 Munich, Germany
Worked as a part-time software engineer and for some time as a research intern. Areas of work included 3D mapping, projection geometry, image and point-cloud manipulation. Languages/APIs: C++, Python, ROS, OpenCV

Awards and Scholarships

Outstanding Reviewer Award
Wrote Exemplary reviews for manuscripts submitted to CVPR 2022.
High Distinction
Completed master’s degree with CGPA of 1.2/5.0 on German scale (~3.8/4.0 on US scale)
Chancellor’s Silver Medal
Completed bachelor’s degree with second-highest CGPA in graduating class (3.93/4.0)
Excellence Scholarship
Partial tuition fee waiver on acccount of >3.6 CGPA throughout all semesters of bachelor’s degree

Contact