Xinnan (Peter) DU | 杜信男
Strategist at Goldman Sachs | CMU MCDS graduate
Contact Info:
Email: xinnand@alumni.cmu.edu
Checkout some of my works:
Self-supervised learning | Roborace self driving car | Visualizing Multi-dimensional Data | Analyzing bitcoin transacitons
About
I'm currenly working at Goldman Sachs as a Equity Derivatives Strat where we build automated pricing and trading systems for structured products. Prior to that I obtained my master's degree from CMU and undergraduate degree from HKUST. I also previously interned at Nvidia and Tencent where I worked on deep learning perception models.
Education
Carnegie Mellon University
Master of Computational Data Science (MCDS)
School of Computer Science, GPA: 3.92
2019 Aug.
2020 Dec.
The Hong Kong University of Science and Technology
Double Major in Mathematics
CGA 3.882 out of 4.3, Computer Science GPA: 4.094/4.3
2014 Sept.
2018 Jun.
Georgia Institute of Technology
Computer Science
GPA 3.8 out of 4.0
2016 Aug.
2016 Dec.
Experience
Jan 2021 - Present
Goldman Sachs
Hong Kong SAR
Automated pricing and trading system for structured products.
May 2020 - Aug 2020
Nvidia, AI infrastructure team
Santa Clara, CA
Self-supervised learning for object detection/panoptic segmentation in AI Infrastructure team for autonomous vehicles
July 2018 - July 2019
June 2017 - Aug 2017
Goldman Sachs
Hong Kong SAR
High-volume post execution trade processing system
Dec 2017 - Feb 2018
Tencent, Youtu Lab
Shenzhen, China
Conducted research on object detection. Implemented and tested state of the art one-stage object detection algorithm (YoloV2, F-SSD, D-SSD) with various techniques such as Focal Loss and Feature Fusion using caffe and tensorflow.
Research & Projects
Boosting Supervised Learning Performance with Co-training
Xinnan Du, William Zhang, Jose M. Alvarez
IV 2021 Paper 2021 IEEE Intelligent Vehicles Symposium
Self-superised learning for deep learning perception models
Source: [pdf]
Roborace (Self Driving Race Car)
Project Built a reinforcement learning training and evaluation framework base on roborace base layer.
The training framework is able to perform distributed training and a dedicated dashboard is built to visualize the training process.
Source: [webpage] [news] [report]
BitExTract: Interactive Visualization for Extracting Bitcoin Exchange Intelligence
Xuanwu Yue, Xinhuan Shu, Xinyu Zhu, Xinnan Du, Zheqing Yu, Dimitrios Papadopoulos, Siyuan Liu
TVCG Paper IEEE Transactions on Visualization and Computer Graphics, 2018
Source: [pdf] [slides] [video] Best FYP Award 2017-2018
Mining and Visualizing Yelp Dataset
Project Built neural network classifier for Health Index of resaurants. Sentiment analysis on large-scale review text using Spark and Map/Reduce on the cloud. Built an interactive online recommendation system for users to visualize local restaurants based on temporal summaries
Source: [github] Champion of Healthcare in HackUST 2017
Awards and Honors
Nov. 2018 | Graduate with first class honors |
Apr. 2018 | Outstanding Graduate Student for Class of 2018, 80/2300 students |
Jun. 2017 | HKSAR Government Scholarship Fund - Reaching Out Award |
Feb. 2016 | HKUST Alumni Endowment Fund & Hung Choh Jan Fong Bursaries/Scholarships |
2015 - 2018 | HKUST Scholarship of Continuing Undergraduate Student |
Multiple Semesters | Dean’s List |
Welcome to Xinnan's homepage
Please view it on a laptop, thanks!