Jiannan Lu
Data Science Manager
Redmond, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolJiannan Lu is a Data Science Manager at Apple with 12 years of experience building experimentation and ML systems for large-scale products, currently leading Siri A/B experimentation. He combines a Ph.D. in Statistics from Harvard with hands-on engineering, having contributed backend and ML code to prominent open-source projects like FacebookResearch's ViLBERT multi-task and PyTorch Faster R-CNN implementations. At Microsoft he helped power the Experimentation Platform, and at Apple he moved from data scientist to manager while continuing to bridge research and production. His background in algorithmic trading and ad measurement shows a strong foundation in causal inference and high-frequency decisioning. Comfortable writing production-ready data loaders, multi-task model integrations, and training APIs, he brings both rigorous statistical thinking and practical software craftsmanship to product experiments. Based in Redmond, he blends academic depth with product-focused delivery and a track record of shipping reproducible, scalable ML infrastructure.
12 years of coding experience
7 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Statistics, Doctor of Philosophy (Ph.D.), Statistics at Harvard UniversityBachelor of Science (B.Sc.), Mathematics and Physics, Bachelor of Science (B.Sc.), Mathematics and Physics at Tsinghua University
English, Chinese
Github Skills (20)
pytorch10
preprocessing10
python10
load-data10
evaluation10
preprocess10
machine-learning10
data-preprocessing10
mask-rcnn10
data-loading10
eval10
dataprep10
deep-learning10
bert10
model-testing10
Programming languages (6)
Github contributions (5)
facebookresearch/vilbert-multi-task
Apr 2019 - May 2020
Multi Task Vision and Language
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:197 commits, 2 PRs, 3 comments in 1 year 1 month
Contributions summary:Jiannan's contributions primarily involve the implementation of data loading, processing, and model integration functionalities within the `facebookresearch/vilbert-multi-task` repository. They implemented a `ConceptCapLoader` for loading data from an LMDB file, demonstrating an understanding of data handling. Furthermore, their work extends to the development and integration of a multi-modal BERT model for vision and language tasks, indicating an understanding of machine learning techniques. They also designed and implemented a system for training models across multiple tasks which involved creating an API for integrating components.
multi-taskvisiondeep-learningcomputer-visionmachine-learning
jwyang/faster-rcnn.pytorch
Aug 2017 - Apr 2018
A faster pytorch implementation of faster r-cnn
Role in this project:Back-end Developer
Back-end DeveloperContributions:125 commits, 1 PR, 1 push in 8 months
Contributions summary:Jiannan's contributions focused on initializing the project and setting up the foundational elements for Faster R-CNN implementation using PyTorch. They created and modified several files, including training scripts, configuration files, and utility modules. These changes involved setting up the initial structure, and establishing paths and configurations for the project.
pytorchr-cnnfasterpytorch-implementationfaster-rcnn
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.