Close

Yutao Zhou

Software Development Engineer at Amazon

M.S. at Columbia University

B.S. at University of California, Santa Barbara

Download Resume

SKILLS & PACKAGES

ABOUT ME

I interned at Amazon in summer 2023 and rejoined Amazon as a full time SDE in Feb 2024. I aim to bring a blend of theoretical knowledge and practical experience to contribute effectively to the company's success.
As a recent graduate of Columbia University, where I earned a Master of Science degree, I am dedicated to leveraging my proficiency in technologies and software development to tackle real-world problems. My educational journey, rooted in a Bachelor of Science from UC Santa Barbara, has honed my analytical thinking and problem-solving abilities.

WORK EXPERIENCE

Amazon Web Service

SDE - Santa Clara, CA

  • Working in a team that changes the way Amazon scales up and tears down its data center networks. We help NDEs with automate as much of their process as possible. Enable Amazon data center capacity(not only AWS) to be elastic.

    • Amazon Web Service

    SDE Intern - Santa Clara, CA

  • Created a Smart Reboot and Host Monitoring System that checks the status of our devices globally and reboots eligible devices with adjustable speed and settings. Redundant checks on host health(e.g. BMC, VPC status, etc.) and location health are performed to ensure the reboot will never influence our services and the system is distributed to 10K+ production devices.
  • Increased devices' health and reduced DevOps labor requirements. All hosts would be patched or updated every 14 days.
  • Completed the project entirely independently with the entire software development cycle from design review, implementation, unit test, integration test, Code review, and monitoring pipeline and deployment to production.
  • Write unit tests using pytest and unittest framework for Python Packages. Write unit test using jtest for Java packages. Achieved total unit test coverage of more than 85% of my code in more than 5 different packages written in Python or Java.
  • Find bugs from other teams and coordinate with their SDEs to fix them. Also, fulfill needs from other teams for shared packages.

    • Deepchem Co., Ltd.

    Python Intern - Beijing, China

  • Designed and built calculation task distribution systems. Distributing calculation jobs from distribution server to different calculation servers (Group project, 4 people in total (including one manager)).
  • Communicated and collaborated with front-end, and other co-walkers to create web-based platform. Represented team to communicate with manage Finished building in 1 month.
  • Checked job status on platform and handled manual stop from user with GET. Handling exceptional cases e.g. distribution server offline. Stress tested on all 4 calculation servers.
  • Checked front-end job status and submitted log content from calculation to distribution server in real-time with GET and POST. Zip needed calculation results and uploaded files to distribution server with POST (up to 10 jobs in real life).
  • Increased overall calculation efficiency by 50% - 200% (By keeping calculation servers busy on none business hours).
  • Created algorithms to find missing data set in database from id queries CSV. Data filtering and aligning. Extract 3D Cartesian coordinate and get SMILE with Pybel(OpenBabel) python package.
  • Constructed and maintain SQL database. Extract data from XYZ file, CSV file, and converted SMILE and insert it into SQL database(including checking repeating data in database)(Individual project).
  • Developed an algorithm to audit two-way connections between PC and lab equipment automatically with SOCKET (Individual project finished in 1 day).

    • Beijing MeiQuan Science and Technology Development Co., Ltd.

    Part-time Software Engineer and Administrative Assistant - Remote, USA / Beijing, China

  • Visualize customer list on interactive map. Click here to see the map. Geocoding with Google Map API and saved to JSON to save API quotes.
  • Read the products’ CAD structure drawings and specifications, make optimization suggestions, participate in new products development, and introduce product features to potential clients.
  • Produce promotional materials remotely, such as the company’s and products’ brochures, using PowerPoint and LaTeX.
  • Discuss project alternatives with the customer including but not limited to problem-solving.
  • Hands-on experience with product production and shipping.

    • Research Institute of Nanjing Runnan Medical Electronics Co., Ltd.

    Part-time Assistant - Remote, USA

  • Collected clinical data in Python and drew signal waveforms, such as electrocardiogram (ECG), electromechanical film ballistocardiogram (BCG), LC BCG, etc.
  • Identified the signal peak through local maximization, calculated the peak distance, and plotted to extract the beat-by-beat cardiac cycle of ECG signal
  • Performed the extraction of the beat-by-beat cardiac cycle of any BCG signal with or without the synchronous ECG signal reference

    • UCSB Physics Circus

    Core member - UCSB, Santa Barbara, CA

  • Deliver intriguing physics lectures to local elementary schools via Zoom
  • Co-designed the Balloon Thermal Experiment and created the featured homepage on Home Experiments Website: https://circus.physics.ucsb.edu/home-experiments/
  • Get reported several times by The Current, UCSB’s official magazine

    • UCSB College of Letter and Science

    Academic peer advisor - UCSB, Santa Barbara, CA

  • Advise peer students on educational plans and provide suggestions on course selection
  • Deal with Qless, a virtual walk-in platform, deal with progress check, petition, explain paper work process, explain policy at College of Letter and Science, and deal with any student concern especially at pandemic time.
  • Deal with academic appointment. Provide detailed academic plan and progress check. Explain potential academic action and consequences.

    • CITI Bank

    YTP Paticipant - Beijing, China

  • Introduce finince product to new custumers. Got awarded with best custumer survice price
  • Listen to morning report, learned a lot of finicial knowledge. Present presentation about group project.

    • EDUCATION

    Columbia University: Fu Foundation School of Engineering and Applied Science

    New York, NY: Sep 2022 – Dec 2023

    Master of Science in Electrical Engineering

    University of California – Santa Barbara: College of Letter and Science

    Santa Barbara, CA: Sep 2018 – Dec 2021

    Bachelor of Science in Physics

    PERSONAL PROJECTS

    JavaScript & React web application series

    Indipendent Projects

  • Link to GitHub repostory for all of the code:Click Here.
  • A fancy Weather App that could change units by clicking on the temperature. Also, the background would change with the weather. Including getting user location and search functions. Build with pure JavaScript, CSS, and HTML. Link to the web application: Weather App.
  • This is a Todo App that would store a to-do list on local storage. It comes with great animations with tons of details. Build with pure JavaScript, CSS, and HTML. Link to the web application: Todo App.
  • This is an online calculator. Build with React. Link to the web application: Calculator App.
  • This is a countdown app. Build with pure JavaScript, CSS, and HTML. Link to the web application: Count Down App.

    • NLP Model's Comparison in Key Phrase Extraction Web Application

    Coorperate Projects

  • Highlight the outcome of different supervised, unsupervised, and bert-unsupervised ML models' key phrases output in the business domain. Adjustable setting with customized natural language input. Total runtime under 1 second in general.

    • Used Car Data Visualization Web Application

    Idependent Projects

  • Link to GitHub repository for all of the code:Click Here.
  • Build with streamlit, dealing large data sets (365K data points) with Desk, Pandas, and NumPy for data filtering and cache data.
  • Visualize data with scatter plot on the heat map (with more than 100 selectable base maps), pie chart, scatter plot with the trend line, with packages e.g. plotly, leafmap, pydeck.
  • Added VIN lookup function with Get from NHTSA (National Highway Traffic Safety Administration)'s API.
  • Designed AI key phrase extraction from listing description with spacy, and visualization with wordcloud with VIN query results (VIN query, key phrase generation, and word cloud should take less than 5 seconds, usually 2 seconds).
  • Implemented geocoding and filtering data with user input distance from user query location with geoencoder in GeoPy (Entire query should take 3 seconds depending on setting usually less than 0.5 seconds).
  • Added Login page with the cookie. Hosting web application on a personal server with domain re-direction (stopped hosting because of the budget problem).
  • Tested google map API with REQUEST for geocoding but abandoned for the cost ($5 per 1000 query).
  • Deployed website to a google cloud platform with docker and yaml files but abandoned it for a daily limit of outgoing internet traffic for app engine with the free account.

    • ACADEMIC PROJECTS

    Concert Buddy

    Course Project: COMS 6156 Cloud Computing

  • Implement microservice architecture, each microservice is a REST API that supports CRUD.
  • Utilized Spring Boot in Java to develop the Concert Microservice deployed on Google App Engine, integrating PostgreSQL hosted on GCP Cloud SQL.
  • Integrated TicketMaster API within the Concert Microservice to synchronize and manage concert information.
  • Implemented Continuous Integration/Continuous Deployment (CI/CD) using GCP Cloud Build, automating build and deployment processes triggered by new commits to the repository.
  • Configured AWS API Gateway and aggregator to route requests across distinct microservices.
  • Employed AWS S3 and CloudFront for Content Delivery Network (CDN), ensuring low-latency delivery of front-end UI websites.
  • Implemented Single Sign-On (SSO) functionality with Google accounts to enhance the user experience.
  • User microservice using Spring Boot, deployed on EC2 using Docker, and utilized PostgreSQL hosted on AWS RDS.
  • Finder microservice with Spring Boot, deployed on EC2, and employed DynamoDB for data storage.
  • Link to GitHub repository for all of the code: Click Here.
  • Click Here to see a demo of the project on YouTube.

    • NYC Subway Traffic Analysis

    Course Project: ELEN 6889 Large Scale Stream Data Analysis

  • Full stack RESTful web application that displays the entry and exit of each subway station on an interactive map.
  • Write Frontend JavaScript that would let the user choose a different time with a slider. Then, corresponding data will be visualized on the map. A ranking of stations with the top 10 throughput at the selected time will be displayed on the side
  • All data presented in the interactive map are fetched in real time from the backend REST API written with Python Flask.
  • The data was batched data from the MTA website and pre-processed with Spark.
  • Link to GitHub repository for all of the code: Click Here.
  • Click Here to see a demo of the project on YouTube.

    • MBTI Personality Analysis and Prediction

    Coorperate Course Project: EECS 6893 Big Data Analysis

  • Link to GitHub repository for all of the code: Click Here.
  • Created a Web application that could use users' Tweets to predict their personalities.
  • Trained a machine learning model using the MBTI personality dataset.
  • We used Flask as the backend and HTML, and CSS as the front end. When a user enters their username we would fetch the user's Tweets using Twitter API(tweepy). Then we would process fetched data and use the pre-trained model to make predictions. Then we would present corresponding results to the user.
  • For details, please check out my final report here.
  • Click Here to view our final presentation on YouTube.

    • Adaptive Bitrate Video Streaming

    Individual Course Project: CSEE 4119 Computer Network

  • The goal of this project is to make all users able to play video smoothly with changing internet throughput.
  • Create a proxy that would be listening(as a server) customer's request(user using a browser), continuously calculating the customer's throughput and requesting video chunks with a suitable bitrate(as a customer) from the video server.
  • Use TCP welcome socket, connection socket, and threading to handle multiple users' streaming at the same time.
  • Note: The course policy does not allow open source for my code. But, I do have all of the code on a private Github Repository.
  • Click Here to see a demo of my proxy on YouTube.

    • My Own internet

    Individual Course Project: CSEE 4119 Computer Network

  • Configure OSPF and iBGP to connect 8 routers and 6 hosts in my Autonomous System.
  • Configure eBGP to perform different routing policies for inter-AS connection with my provider, customer, and peers. e.g. Achieved no valley routing. Achieved preferred customer routing(preference in this order: customer, peer, provider). Inbound traffic engineering: prefers traffic coming from one link of a provider(that has multiple links). Guide traffic to prefer coming from one provider over others.
  • Successfully fetched data across our internet(with working policy) formed with my classmates.
  • Note: The course policy does not allow open source for my code.

    • Swim Transformer Paper Reimplimentation

    Individual Course Project: ECBM 4040 Deep Learning and Neural Network

  • Replicated the entire Swin Transformer architecture.
  • Tested my model on CIFAR-100 and achieved an accuracy of 46.7%(only have access to very limited computational resources as a result could only train on a very small dataset compared to the original paper).
  • For details pleas check out my final report here.
  • Note: The course policy does not allow open source for my code. But, I do have all of the code on a private Github Repository.

    • GraphQL API for OpenSea Smart Contract

    Course Project: ELEN 6883 Blockchain Technology

  • Write SubGraph(an GraphQL API) that eable user to interact with OpenSea Smart Contract without need to understanding technical details under the hood.
  • User could query information, listing, transfer, and betting NFT on OpenSea Smart Contract with GraphQL query syntext.

    • Introduction to Scientific Computing Laboratories

    PHYS 129L Course Projects

  • Using Python to do: string processing, error handling, inheritance, and derived Julian Day based on the input
  • Calculated the complex plane, drew a fractal image, and plotted a 3D image of a rotatable airy disk
  • Practiced Discrete Fourier Transform, Fourier analysis algorithm, and drawing method
  • Exercised Gaussian approximations, integral algorithm, and Monte Carlo simulation

    • Auroral Morphology Classification Based on Unsupervised Clustering

    Advisor: Prof. Ziming Zou (National Space Science Center, CAS)

  • Familiarized with the principle of unsupervised clustering algorithm and the process details of KNN algorithm and K-means algorithm
  • Learned about calling the underlying algorithm in the SkLearn machine learning library
  • Consulted literature on the morphological categories of aurora observation images in the whole sky and collected the data sources of aurora images
  • Programmed to realize morphological clustering of aurora images

    • Get in Touch

  • Email: yutaozhoucolumbia@gmail.com
  • Tel: +1 805-637-1617
  • School Email: yz4359@columbia.edu
  • Tel: +86 13520759678
  • Address: 3131 Homestead Rd, Santa Clara, CA 95051