Chat
Randolf Uy
Software Engineer | CS @ SFSU
About Me
Hi! I am Randolf, a December 2024 graduate from San Francisco State University. I've always had an interest in building things and solving challenges. Currently, I work at the intersection of DevOps, Data Engineering, and Infrastructure Engineering.
I strive to create impactful software. Knowing that I can contribute to something meaningful motivates me. Throughout my life, whether in group projects, extracurricular activities, or professional settings, teamwork and strong leadership have been essential to my growth.
Beyond engineering, I aim to develop as a well-rounded person. I enjoy working out, playing basketball, watching K-dramas, and spending time with loved ones.
My current goals include contributing to Apache Foundation open-source projects, building tools that benefit others, and—most importantly—finally learning how to exit Vim correctly!
Education
San Francisco State University
I earned a Bachelor of Science in Computer Science. My most notable coursework includes: Software Engineering, Software Development, Operating Systems, Databases, Artificial Intelligence, Deep Learning, Data Structures, and Algorithm Analysis.
Experience
Head of Leadership and Development - Data Science Society @ SFSU
Club infrastructure and personnel development within the organization. Initiatives with SQL and information retrieval.
Technology used: SQL (SQLite), PySpark, PyTorch, Python
Data Engineer / Software Engineer Intern - PricewaterhouseCoopers (PwC)
Part of the data engineering team. Worked on Kubernetes / Artifactory logging, API development for data pipeline, and addressing Azure security risks.
Technology used: Python (FastAPI), Azure Cloud, Kubernetes, Docker, Airflow
Technology Consultant Start Intern - PricewaterhouseCoopers (PwC)
Data analytics and competitor analysis for a client. Placed top 6 across all pods nationwide.
Technology used: Python (BeautifulSoup, Selenium), Excel, Tableau
Projects
Gopher-DB
An open-source, lightweight database that supports concurrent, ACID-compliant transactions.
Technology stack: Go
Responsibilities:
- Architecting the storage engine for the database, from the pager layer to the B+ tree index.
- Created documentation outlining design and engineering decisions for long-term maintainability.
ChatFlow
AI agent created for the purpose of answering questions anyone may have on me with specficiations and no generalizations. Can be tailored to be used for someone else.
Technology stack: Python, Pinecone DB, Openai API, AWS
Responsibilities:
- Aggregated pdf data about myself and created parser to extract data in an efficient manner
- Created API interface to embed and load parsed data into pinecone's vector database. Engineered agent model to reference documents from database in order to answer questions with context.
- Deployed on AWS Linux using docker container. Configured Nginx as a reverse proxy to prevent unauthorized API / Websocket access.
- Utilized redis as a rate limiter in order to prevent excess requests to the websocket interface.
Yelp Reviewer
Classifcation model based on transformer architecture used to predict star rating of various restuarants reviews using the traditional 1 through 5 model.
Technology stack: Pyspark, Pytorch, Huggingface, FastAPI
Responsibilities:
- Preprocessed ~5 gigabytes of JSON yelp data by leveraging Pysparks big data analytics decreasing script preprocessing time from ~20 min to ~5 min locally.
- Architectured Pytorch encoder classifcation model to accuractly (80%) predict yelp review
- Trained model by leveraging google collab's ability to use a100 GPU to accelerate training speed by 5x (roughly 5 hours per epoch to 1 hour per epoch)
- Provided performance metric against traditional regression models and NLTK models seeing 2x increase in accuracy
File Allocation Table (FAT-32) Filesystem
Linux Ubuntu Filesystem capable of creating/deleting/moving files, creating directories, and navigation. Capable of up to 32 GB of persistent data.
Technology stack: C, Linux, Systems Administration
Responsibilities:
- Engineered deallocation of memory by using the freespace chain minimizing data fragmentation. This lead to accurate read / write operations which used 95% of space efficiently.
- Developed directory functions and documentation for file system traversal through using a current working directory ensuring proper traversal 100% of the time.
- Solved technical problems regarding write operations for copying a file by using a buffer based algorithm allowing for 100% accurate file copy operations.
Team-Match
Team organization app centered around community and events for respective hobbies.
Technology stack: React.js, Express.js, Typescript, MySQL
Responsibilities:
- Collaborated with frontend team on API accessibility and security issues reducing risks by 80%.
- Designed backend architecture for interaction with database increasing poductivity by 75% across team.
- Unit and end to end testing for products landed ~95% API reliability.
- Responsible for creating data models and backend infrastructure.
- Deployed server on AWS and responsible for database migrations.