CV

Alternatively, open the PDF version here.

Curriculum Vitae

Updated January 2026. This page includes an embedded PDF and a structured HTML version for quick scanning.

Contact

+44 7849 244 069

tonyngjichun@gmail.com

LinkedIn

GitHub

Google Scholar

Summary

Scientist at Meta AI Research with a PhD in Electrical and Electronic Engineering from Imperial College London. Focused on diffusion models, visual localization, and privacy-preserving perception systems.

Experience

AI Research Scientist
Meta 路 London 路 Aug 2024 - Present
  • Image, video, and audio diffusion models to optimize ad performance across Meta platforms.
  • Data curation pipelines from pretraining (millions of assets) to post-training (thousands of high-quality samples).
  • Evaluation guidelines for generation quality, deployed at scale with annotators.
  • Boosted advertiser adoption by 40% on a $1M+ daily revenue stream via image generation models.
Research Engineer
Synthesia 路 London 路 Feb 2023 - Aug 2024
  • Controllable video diffusion models for AI dubbing on avatars.
  • End-to-end R&D and deployment of AvatarBuilder in Synthesia STUDIO.
  • System design, engineering, and deployment for AI video generation at scale.
Research Intern
Meta 路 London 路 Jun 2022 - Oct 2022
  • Led a research agenda in SceneScript for multimodal text-to-3D understanding.
  • Built an internal API and demo for a scene generation pipeline.
Research Intern
Facebook (Reality Labs) 路 Redmond, WA 路 May 2021 - Nov 2021
  • Led CVPR 2022 research on privacy-preserving ML using adversarial learning.
  • Produced research code compatible with Meta infrastructure.
Teaching Assistant
Imperial College London 路 London 路 Oct 2018 - Aug 2023
  • Courses: Machine Learning (Year 3), Pattern Recognition (Year 4), Deep Learning (Year 4).
  • Marking, feedback, and tutorials.
Visiting Student
Scape Technologies 路 London 路 Jun 2019 - Dec 2019
  • Job-shadowed Dr. Vassileios Balntas and team.
  • Assisted integration of image retrieval modules into localization pipeline.
Summer Intern
Rolls-Royce plc 路 Derby, UK 路 Jun 2016 - Sep 2016
  • Managed a 7-person team to resolve a supply chain disruption (~GBP 120K impact).
  • Analysed 650k+ entries to assess replacement risks and costs.

Education

PhD, Electrical and Electronic Engineering
Imperial College London 路 Sep 2018 - Aug 2023
  • Research topic: Visual localization for AR/VR using deep learning and geometry.
  • Supervisor: Prof. Krystian Mikolajczyk, Co-supervisor: Dr. Vassileios Balntas.
  • Scholarship: Departmental Scholarship for Overseas PhD Candidates.
MEng, Aeronautical Engineering
Imperial College London 路 Oct 2014 - Jun 2018
  • Thesis: Prediction Model for Gravity Currents. Grade: 80/100.
  • Classification: First Class Honours (1:1), Average Score 78%.
  • Awards: Dean's List 2015 & 2016, Aeronautics Scholars.

Publication Highlights

  • NinjaDesc: Content-Concealing Visual Descriptors via Adversarial Learning (CVPR 2022).
  • OoD-Pose: Camera Pose Regression from Out-of-Distribution Synthetic Views (3DV 2022).
  • SOLAR: Second-Order Loss and Attention for Image Retrieval (ECCV 2020).
  • HyNet: Learning Local Descriptor with Hybrid Similarity Measure and Triplet Loss (NeurIPS 2020).
  • D2D: Keypoint Extraction with Describe to Detect Approach (ACCV 2020).

Service & Leadership

Reviewer
CVPR 2023
  • Outstanding Reviewer.
Reviewer
NeurIPS 2021 & 2022
  • Outstanding Reviewer Award (top 8% in 2021).
Reviewer
CVPR 2022, CVPR 2025, ICLR 2022, ECCV 2022
  • Conference reviewer.
Invited Speaker
Image Matching Challenge (CVPR 2020)
  • Lead presenter for a top entry in Local Features and Beyond.
Leader / Arranger / Singer
The Mockingbird (Acapella group) 路 London 路 Oct 2014 - Mar 2017
  • Founded by students and music enthusiasts from Hong Kong.
  • Led concerts with hundreds of attendees; bass singer and song arranger.
Associate
Imperial College Consultancy Society 路 London 路 Oct 2016 - May 2017
  • Hosted and prepared materials for weekly case-study sessions.

Skills & Interests

Computing
  • High: Python, PyTorch
  • Medium: Docker, k8s / Kubeflow
  • Basic: JavaScript, C++
Languages
  • Native/Fluent: English, Cantonese, Mandarin
  • Conversational: Korean, German (B1+)
  • Basic: Japanese
Interests
  • Singing, travelling, lifting weights, tennis, Formula 1, Age of Empires II