Harrison Stark
I am a forward-thinking computer science enthusiast deeply committed to innovation and pushing technological boundaries. My passion for continuous learning is evident through hands-on experience and cutting-edge projects. My values extend beyond technology, reflecting a holistic educational background and my relentless pursuit of knowledge.
For information about my hobbies, click here.
Education
Bachelor of Science in Computer Science, Minor in Mathematics - 3.78 GPA
(August 2021 - December 2023)
University of Florida
Associate of Arts - 4.0 GPA
(January 2019 - May 2021)
Palm Beach State University
High School Diploma - 3.94 GPA
(August 2017 - May 2021)
Suncoast Community High School
Experience
Qualtrics - Software Engineer 1 (January 2024 - Current)
- Accelerated Machine Learning Infrastructure:
- Managed and implemented features for a low-latency, high-traffic API gateway for company-wide machine learning inference supporting billions of monthly inferences.
- Integrated with cutting-edge models, implementing cost-tracking mechanisms, and ensuring real-time metrics with effective visualizations.
- Unified API Developments:
- Created a comprehensive embedding API from scratch.
- Integrated with Azure OpenAI, AWS Bedrock, and self-hosted AWS SageMaker models.
- Spearheaded the general availability readiness of our LLM API including robust reporting and comprehensive scale testing protocols and documentation.
- Internal Tool Enhancements:
- Upgraded our internal LLM playground with significant new features such as chat history and CI/CD.
- Guided an intern through project development, ensuring platform compliance with enterprise-grade standards for the internal LLM playground.
- Led a team in the development of an internal LLM-powered Slack bot designed to triage and resolve tickets by suggesting relevant information from various sources, improving ticket resolution times and enhancing ticket experience.
- Cost Tracking and Data Collection Systems:
- Designed and deployed dashboards for extensive cost tracking across models, utilizing advanced data analytics tools.
- Automated critical data collection processes for Monthly Engineering Review (MER) to reduce manual overhead and enhance accuracy, efficiency, and visibility across brands and inferences.
- Implemented a company-wide AWS policy enforcing billing code tagging across all roles to ensure precise cost tracking for SageMaker resources.
- Operational Readiness and Documentation:
- Led the creation of extensive on-call runbooks, a standard documentation repository for our team, and standard operating procedures.
- Improved operational readiness and ensured smooth knowledge transfer.
- Cross-Team Collaboration and Support:
- Facilitated cross-functional meetings.
- Presented design documents and led initiatives to streamline processes, optimize resource allocation, and address team-wide challenges.
Qualtrics - Software Engineer (Part-Time) (October 2023 - December 2023)
- Production Readiness:
- Prepared the Quack-end Developer project for production, ensuring launch readiness and internal access for all employees.
- Security and Maintenance:
- Migrated high volume automations from Python 2 to Python 3, enhancing security posture.
- Developed comprehensive unit and integration tests and thorough documentation for ease of future development.
Qualtrics - Software Engineer Intern (June 2023 - August 2023)
- LTI 1.3 Integration:
- Designed and developed the Gradebook Passback feature for the Qualtrics LTI 1.3 integration using Node.js, facilitating seamless grade data synchronization from Qualtrics surveys.
- Permissions Feature Development:
- Proposed and integrated a permissions feature, significantly expanding the potential customer base.
- Launch Readiness Preparation:
- Executed thorough testing, review, and completion of the launch readiness checklist.
- Enabled a successful early access launch.
Undergraduate Research - Independent Researcher - (Link to Repository) (January 2023 - May 2023)
- Authored a research paper exploring the use of YOLOv8, a state-of-the-art object detection algorithm, to improve real-time traffic signal operations at a complex intersection in Gainesville, Florida.
- Suggested an implementation plan involving cameras connected to a central processing unit to monitor traffic and pedestrian activity.
- Proposed signal timing adjustments to improve safety and efficiency up to 10%.
Projects
Legame (June 2024 - Current)
- Developed a farming simulation game using C# in Godot, focusing on achieving a maximum farm size, paying off loans, and becoming self-sufficient to retire in the shortest possible time.
- Implemented mechanics for planting and harvesting various legumes, each with unique growth patterns and special abilities, to create a dynamic and engaging gameplay experience.
- Designed intricate gameplay systems including inventory management, item drops, legume hybridization, and upgrade paths to provide players with complex strategic choices.
Stark's Industry Mentorship (February 2024 - Current)
- Mentored students at the University of Florida through tutoring, project management, and mock interviews to enhance their understanding of the tech industry.
- Led students in industry-standard project lifecycle development, emphasizing agile methodologies, version control best practices, and thorough documentation.
- Guided a student in developing a full-stack banking application, providing practical experience in real-world software development, including concurrency handling, UI and API design choices, and database management.
Harrison Stark's (My) Website - (Link to Project) (November 2023 - Current)
- Developed a personal website using HTML, CSS, and JavaScript to effectively present information about my background, skills, and projects.
- Designed and implemented a user-friendly interface that highlights my professional experience and portfolio.
- Leveraged the project to enhance my understanding of fundamental web development concepts and improve my proficiency in frontend technologies.
MAISTRO - (Link) - (Frontend) - (Backend) (Fall 2023)
- Led a team of four in developing an AI DJ using JavaScript and Python, adhering to enterprise software engineering standards.
- Integrated the Spotify SDK to create a seamless user experience with a unique song recommendation algorithm utilizing the ChatGPT and Spotify APIs and a custom NLP model for emotion and valence extraction.
Quack-end Developer (Summer 2023)
- Led a team of three in a two-day AI Hackathon to develop an LLM-based rubber duck debugger.
- Assisted in creating the frontend utilizing React.js and led development of the backend powered by FastAPI.
- Implemented text-to-speech and speech-to-text using Amazon Polly and Amazon Transcribe, respectively, and used AWS S3 Buckets for rapid input and output.
Only Cats - (Link to Project) (Fall 2022)
- Developed a video processing program with a partner to remove video frames not containing cats in Python.
- Utilized a CNN for cat detection in videos, implementing an enhanced EfficientNetB0 architecture.
- Accelerated unit testing by up to 99% using single images in place of videos.
Traffic Accident Heatmap - (Link to Project) (Spring 2022)
- Worked in a group of three to create a C++ program to display traffic accidents using the Google Maps API.
- Adapted dataset of nearly three million traffic accidents from Kaggle which included location and severity.
- Improved time efficiency of click response and sort by over 90% by utilizing attributes of heaps.
Assembly Simulator - (Link to Project) (Spring 2022)
- Created an assembly program simulator in C++ for ARM programming.
- Implemented functionality of 16 ARM commands, along with branching and a program counter.
- Included ability to produce disassembly and simulation files from binary input.
Pokémon Pokédex - (Link to Project) (Spring 2021)
- Utilized Java and Swing to create a scrollable GUI Pokédex with information about every Pokémon.
- Implemented a SQL database using JDBC to fetch and store Pokémon and user information devised from a UML diagram.
- Improved program incrementally using client-supplied feedback, success criteria, and a Gantt chart.
Landscape Classifier - (Link to Paper) (Spring 2021)
- Developed a CNN to classify 150x150 RGB images for a research paper: The Effects of Hyperparameter Choices on Convolutional Neural Networks using Python
- Incorporated elements from the following libraries: TensorFlow, NumPy, and Matplotlib.
- Determined the most accurate activation function-kernel size pair to be Leaky ReLU and (3, 3).