Tai Vu

I’m an AI enthusiast, an innovative thinker, a risk taker, and a team player.

🏅 AWARDS

  • Bronze Medal in the 2016 International Mathematical Olympiad.
  • Gold Medal in the 2016 Vietnamese Mathematical Olympiad (Rank #1).
  • Gold Medal in the 2016 American Mathematics Competition.
  • Gold Medal in the 2015 International Mathematics Local Tournament.
  • Silver Medal in the 2015 Vietnamese Mathematical Olympiad (Rank #2).
  • Gold Medal in the 2015 and 2014 Regional Mathematics Competition.
  • Gold Medal in the 2015 and 2014 Nam Dinh Province Mathematics Competition.
  • Gold Medal in the 2015 High School for Gifted Students Olympiad.
  • Gold Medal in the 2014 Hanoi Open Mathematics Competition.
  • Scholarship of the National Program for the Development of Mathematics in Vietnam.

🖥️ WORK EXPERIENCE

Meta Platforms

Senior Machine Learning Engineer

July 2022 - Present • Menlo Park, CA, USA

  • #1 impact driver to organizational topline metrics, directly contributing to 90-120% of revenue and session targets and 40-50% of DAU goals through end-to-end ownership of high-leverage ML solutions. Recognized with Exceeded Expectations+ ratings over multiple halves
  • Spearheaded large-scale, cross-organizational initiatives spanning 20+ engineers across infrastructure, product, and modeling teams. Independently defined strategies, aligned stakeholders, and executed system-wide upgrades that materially advanced Meta’s notifications recommendation system.
  • Pushed the frontier of recommendation modeling by pioneering advanced architectures and optimization strategies. Developed novel techniques including LLM-powered candidate generation and ranking, contextual and long-horizon representation learning, dynamic user interest modeling, causal inference, knowledge distillation, pairwise ranking, and value-aware multi-task learning to capture latent interests and maximize user engagement.
  • Scaled team capability by 3x by proactively scoping and launching new technical roadmaps, mentoring and onboarding 10+ IC4/IC5 engineers, and creating a culture of technical rigor that enabled sustained growth and research-grade innovation.

Meta Platforms

Software Engineering Intern

June 2021 - September 2021 • Menlo Park, CA, USA

  • Worked on the Facebook Commerce Monetization team.
  • Developed large-scale ML models in PyTorch to generate contextual embeddings of users and ads, enhancing semantic understanding and fine-grained personalization in ad retrieval and ranking systems.
  • Designed and deployed representation learning frameworks for user-ad matching across multiple surfaces, significantly improving relevance and driving ad revenue growth. Used Python, SQL, C++, PHP, and Spark.

Meta Platforms

Software Engineering Intern

June 2020 - September 2020 • Menlo Park, CA, USA

  • Worked on the AI Mobile Platform team, contributing to the development of the PyTorch platform.
  • Enhanced latency profiling tools with module-level debugging. Achieved a 5x increase in runtime efficiency.
  • Accelerated Conv1D and channel shuffle operations through low-level kernel optimizations for the PyTorch framework, delivering up to 10x operator-level speedup for on-device speech and NLP models. Used Python, PyTorch, and C++.

🔬 RESEARCH EXPERIENCE

Stanford AI Lab

Undergraduate Researcher

April 2020 - September 2021 • Stanford, CA, USA

  • A member of the Stanford Machine Learning Group. Worked on the AI for Climate Change Bootcamp under Prof. Andrew Ng’s supervision.
  • Built large-scale data pipelines in Python, NumPy, and Pandas to ingest, align, and preprocess satelite imagery and global forest loss driver labels. Implemented advanced data augmentation and stratified sampling techniques to improve signal diversity and model robustness across diverse geographies.
  • Designed and trained deep learning models in PyTorch, including CNNs, LSTMs, and multimodal fusion architectures, to classify forest loss drivers from multi-temporal satelite imagery, achieving 80% classification accuracy and supporting scalable environment monitoring.

Stanford InfoLab

Undergraduate Researcher

February 2020 - April 2020 • Stanford, CA, USA

  • Engineered a high-throughput input pipeline for loading and preprocessing underwater video data, including image normalization, spatial augmentation, and temporal slicing, to support robust model training.
  • Developed Mask R-CNN and U-Net models in TensorFlow to detect, localize, and temporally track coral structures in underwater environments, enabling fine-grained analysis and monitoring of reef health over time.

Computer Science Research Lab

Undergraduate Researcher

September 2019 - December 2019 • Stanford, CA, USA

  • Researched model compression techniques, including structured pruning, regularization-based sparsity, and weight quantization, to reduce the computational footprint of deep convolutional networks.
  • Applied regularization and pruning strategies to ResNet, achieving a 15% reduction in FLOPs with a 2% improvement in accuracy, demonstrating efficient compression without performance trade-offs.


🏫 EDUCATION

Stanford University

B.S., Computer Science (AI Specialization)

Graduated June 2022 • Stanford, CA, USA

  • GPA: 4.18 / 4.0. Graduated with Distinction (Summa Cum Laude).
  • Coursework:
    • Computer Science: Machine Learning, Deep Learning, Natural Language Processing, Large Language Models, Reinforcement Learning, Computer Vision, Natural Language Understanding, Spoken Language Processing, Data Mining, Machine Learning with Graphs, Data Structures & Algorithms, Computer Systems, Object-Oriented Design, Parallel Computing, Networking, Information Retrieval, Web Applications, Databases & Data Systems, Randomized Algorithms.
    • Mathematics: Linear Algebra, Probability & Statistics, Convex Optimization, Real Analysis, Calculus, Graph Theory.
    • Others: Cryptography, Economics, Econometrics, Game Theory, Entrepreneurship, Physics, Electrical Engineering, Scientific Writing.


📕 PUBLICATIONS & PROJECTS

I have completed a number of research studies and programming projects in machine learning, NLP, computer vision, web development, and app development. Below are some of them.

  • FlapAI Bird: Training an Agent to Play Flappy Bird Using Reinforcement Learning Techniques [paper] [github].
  • How Not to Give a FLOP: Combining Regularization and Pruning for Efficient Inference [paper] [github].
  • Privacy Preserving Inference of Personalized Content for Out of Matrix Users [paper] [github].
  • GANime: Generating Anime and Manga Character Drawings from Sketches with Deep Learning [paper] [github].
  • BERT-VQA: Visual Question Answering on Plots [paper] [github].
  • Pixel-Perfect Piloting: Superhuman Control of Pixelcopter via Reinforcement Learning [paper] [github].
  • Beyond the Panels: A Deep Neural Network Approach for Manga Object Detection [paper] [github].
  • Amplifying Emotional Signals: Data-Efficient Deep Learning for Robust Speech Emotion Recognition [paper] [github].
  • From Bayes to BERT: A Comprehensive Benchmark for State-of-the-Art Intent Detection [paper] [github].
  • ConnAIsseur: An AI-driven Recipe Recommendation Website [github].

To learn more about my projects, please visit my portfolio.


🤹🏼 SKILLS

  • Languages: Python, C, C++, JavaScript, TypeScript, Java, SQL, PHP, Assembly, R, MATLAB.
  • Technologies:
    • Artificial Intelligence: PyTorch, TensorFlow, Keras, HuggingFace, Scikit-learn, OpenCV, NLTK, SpaCy, Detectron2, CoreNLP, FastAI, Scikit-image, MLflow, XGBoost, Dlib.
    • Data Analysis & Processing: NumPy, Pandas, Spark, Matplotlib, Seaborn, SciPy, Statsmodels, NetworkX, D3.js, Tableau, Stata.
    • Web Development: HTML, CSS, Sass, React, Node.js, MongoDB, Express.js, Next.js, Django, Material-UI.
    • App Development: React Native.
    • Cloud Computing: Amazon Web Services, Heroku, Google Cloud Platform, Microsoft Azure.
    • Developer Tools: Git, GitHub, Linux, Visual Studio Code, XCode, Android Studio, R Studio.