Zengyongxi Deng

Carnegie Mellon University

Sep. 2022 - Dec. 2023 | Pittsburgh, PA

Master of Information Techonology and Strategy

• Concentration: Software and Networked Systems
• Coursework: Computer Systems, Distributed Systems, DevOps and Continuous Integration, Search Engines, Machine Learning, Info Security & Privacy, Database Systems, Object Oriented Programming

China University of Political Science and Law

Sep. 2017 - Jul. 2021 | Beijing, China

Bachelor of Economics

• GPA: 3.81/4.0, Top 1% Rank
• National Scholarship & Excellent Undergraduate | Exchange Student at UCLA & Nankai University

Search Engine with Learning to Rank

Feb. 2023 - May 2023 | CMU 11642

• Search Engine System: Constructed a thorough search engine with OOP architecture, supporting multiple exact-match and best-match retrieval models, including Ranked Boolean, Unranked Boolean, BM25 and Indri models using Java.
• Machine Learning: Applied Machine Learning method, such as SVM, Coordinate Ascent and ListNet to the constructed search engine, improved the model's Mean Average Precision by 20.5%.
• Pseudo Relevance Feedback: Optimized model parameters and designed structured queries through heuristics and experiments, implemented Pseudo Relevance Feedback to expand and diversify queries, improved the model's Mean Average Precision by 68.7%.

Automated Containerized CI/CD Pipeline

Mar. 2023 - Jun. 2023 | CMU 17646

• Continuous Integration: Designed a Jenkins pipeline, automating code cloning, building, SonarQube based static analysis and deployment of an open-source Spring PetClinic application on a Docker generated Ubuntu container. Monitored the roll-out process using BlueOcean.
• Continuous Delivery: Configured an Ansible playbook, automated application deployment to target web servers.

Computer Systems Labs

Sep. 2022 - Dec. 2022 | CMU 15513

• Cache Simulation: Conducted a customized cache simulator in C, implemented set-associative mapping, LRU replacement policy and write-back & write-allocate policy using doubly linked list.
• Manual Memory Allocation: Optimized memory allocation with implicit and explicit lists, incorporating both LIFO and FIFO queue strategies. Achieved a 34% throughput improvement compared to benchmark via a segregated free list, footer elimination, and minimized block size.
• Proxy Server: Developed a multi-threaded web proxy using Pthreads, parsing & forwarding HTTP requests and thread-safe caching of web objects.

TRACT

Jun. 2023 - Aug. 2023 | San Francisco, CA

Software Development Intern

• AI Chatbot: Devised Sherlock Report AI Chatbot using Flask and Python for Sherlock platform, built MongoDB toolkit enabling LangChain-supported LLM for conversational memory, integrated LlamaIndex query engine responding dynamic user requests, decreased response time by 40% from MongoDB.
• SEO: Conducted a multi-threaded automated AI Blog generator for Search Engine Optimization using OpenAI APIs. Generated 100+ HTML formatted articles through asynio and aiohttp with fine-grained keywords, monitored the blog generating process with tqdm.

Global Resource & Technology Development Inc.

Mar. 2023 - Jun. 2023 | New York, NY

Software Engineer Intern

• Interactive Question Handling: Spearheaded the backend of online education product MathZoos using Python. Designed and implemented user interaction features, including event listening, bookmarking, and hint requesting.
• Customized Practice System: Crafted the personalized daily practice system based on Flask, deployed the project on AWS EC2. Dynamically evaluated user performance on a sliding window leveraging PostgreSQL and SQLAlchemy, leading to a 15% improvement in user retention.

Shining Star Investment Mgmt. Co., Ltd.

Nov. 2020 - Jan. 2021 | Beijing, China

Quantitative Development Intern

• Platform Migration: Migrated & optimized original implementation of RiceQuant-based data extraction, avoided repetitious extractions by setting time data in a separate table that served as surrogate keys, reduced run times by 80%.
• Streamlined Data Automation: Constructed an automated data extraction & processing pipeline in Python using shell scripts for 3045 stocks in Chinese A shares market, utilized HDF5 for persistent cross-platform data storage, reduced data processing time by 75%.
• Technical Report: Analyzed effectiveness of 1000+ factors for stocks selection, cleansed & transformed 10+ volume- price factors using Alphalens, generated factor group statistics, IC, return rate and turnover rate analysis charts with Matplotlib & presented them to senior management with data-driven insights on 3 selected stocks.