Yuqing Qiu

Apple Inc.

Software Engineer Intern

May 2023 - August 2023

• Developed a monitoring and alerting system for ads platform to enable feature observability of downstream logs.
• Engineered a scalable infrastructure using AWS EMR and Docker for efficient data processing to enhance monitoring cadence.
• Streamlined end-to-end process with a CI/CD pipeline and Airflow orchestration to optimize workflow management.
• Implemented automated data quality checks with prompt anomaly alerts to enable model calibration for better performance.

Intel Corporation

Deep Learning Software Engineer Intern

May 2021 - August 2021

[repo]

• Wrote APIs utilizing PySpark framework to support the distributed cluster serving for large-scale recommender systems.
• Optimized DLRM data preprocessing on Twitter dataset by adapting Spark join strategies.
• Trained a recommendation system from a WeChat dataset of over 10 million video feeds to predict user actions.
• Deployed sentimental analysis example on distributed training pipeline and scaled out from single node to big data clusters.

Carnegie Mellon University Parallel Data Lab

Software Engineering Intern (Part-time)

Jan 2023 - May 2023

• Engineered a cost-effective cloud-edge emulating framework for testing cloud resources locally to reduce the operational costs.
• Built a streamlined pipeline for launching pods in Kubernetes with customized network topology and application deployment.
• Deployed a video surveillance application onto emulator with seamless cloud-edge collaboration for iterative model refining.

Serving System for Unpredictable DNN Inference Requests

[Code] [Paper]

• Instructed by Professor Mosharaf Chowdhury.
• Explored 10+ NLP models and synthesized 20+ evaluation datasets integrating 1998 World Cup and Microsoft Azure job traces.
• Deployed the proposed bucketing algorithm and performed data analysis in job completion, SLO satisfaction and resource utilization.

Investigating Unsupervised Learning Methods To Select Fair Samples Of Large Unlabelled Image Datasets

[Poster]

• Instructed by Professor Carol Flannagan.
• Designed the sequential sampling method based on the stream-based active learning to reduce the cost of manual labeling and bias in large unlabeled image datasets.

Parallel Four-Color Map Solver

[Website] [Repo] [Paper]

• Designed and built a parallel map solver for the four-color problem using openMP and performed a detailed analysis of performance characteristics.

3-Way Superscalar R10K-Style Out-of-Order Microprocessor

[Paper] [Slides]

• Designed, built, and synthesized a 3-way scaled processor in Verilog with out-of-order instruction execution, register renaming, prefetcher, GShare BP and BTB to support 32 bits RV32IM ISA.

Distributed and Fault-Tolerant Key/Value Services

• Built a sharded storage system in Golang using consistent hashing to support intensive Get, Put and Append RPC requests.
• Implemented Paxos replication Protocol to tolerate network and server faults, and improve load balance and availability.

Highlight Moments Capture App with Explainable Image Assessment

• Developed and deployed an android app in Kotlin for automatic highlight capture with Django backend and based on SAMP-Net model for aesthetic image scoring.

Online Social Media Platform

• Developed an Instagram-like website with posts, comments, and follows in Python with SQLite3 database and Flask backend and built client-side dynamic pages on the frontend using React.js.

Compiler Construction

• Designed a 64-bit end-to-end compiler in Rust using lexical analysis to support multiple data types, binary arithmetic, conditions, mutable arrays, function declarations and calls, and assembly code generation.

Automatic Image Colorization with CNN and GAN

[Paper] [Code]

• Implement Generative Adversial Networks (GAN) for image colorization on CIFAR-10 dataset.
• Compare it with the classification approach in terms of PSNR and SSIM.