- Develop novel computer vision pipeline using image pre-processing and Vision Transformers to automate MRI vocal tract segmentation, reducing hours of manual labeling (submitted to Acoustical Society of America 2024)- Visualized multilingual LLM interpretations by comparing speech Transformer LLM (Meta AI HuBERT) activations in German vs. English vowels- Benchmarked cost vs. latency of Meta AI SAM model on Nvidia A100, L4, and T4 GPU to determine hardware needed for containerizing and deploying the model in Google Kubernetes Engine

- Designed and implemented LLM Benchmarking tool using NextJS and Google Cloud for performance evaluations between model releases–allowing competitive analysis for Engineering and Solutions Architect teams- Developed CI/CD workflow in Github Actions to automate unit testing with docker containers to streamline development-to-production releases- Performed dual role as Software Engineer and Product Manager–drive MVP feature plans, align stakeholder visions into actionable design docs, and translate ambiguity into concrete engineering tasks

- Teaching Intro C++ cohort of 40 students by leading a weekly lab section- Host weekly office hours and collaborate with the professor to create lessons, assignments, and exams

- Debugged ARM-based C++ firmware in Linux-based environment using GDB and Logic Analyzers- Responsible Engineer for in-house rocket tracking system, managing deliverables for SW, EE, and MechE teams- Produce detailed project and toolchain collateral to simplify onboarding using Github, GNU, and Python

- Spearheaded NLP pipeline using Python to automate speech-to-text translation and summarization using SpaCy LLM and AWS S3 Buckets, resulting in converting a customer from trial to a paid subscription- Enhanced LLM accuracy and reduce token consumption through efficient use of few-shot prompting engineering, resulting in $36,000/year savings in operation cost- Pioneered “Meta-Prompting”, a workflow emulating prompt engineering through recursively calling an LLM to create and score new AI generated prompts- Implemented AWS Relational Database endpoints to query and slice NLP pipeline findings by metadata uploaded during integration, resulting in more scalable and granular customer insights- Overhauled codebase error handling to use centralized Object Oriented architecture between frontend components and backend queries to recover from customer-facing failures- Developed internal guides on Infrastructure as Code (IaC) and AWS Serverless Stack Technology (SST) to streamline onboarding for new engineers

Roles and Accomplishments-Led team to compete in Alabama and Northern California Regionals Championships-Contributed to open source robot navigation library to resolve hardware null-pointer crashes-Developed collateral to drive adoption of AprilTags open source computer vision library -Optimized Blob Detection computer vision algorithm for 97% accuracy object detection under variable lighting

Roles and Accomplishment- Captain for completely student-led team, advanced to NorCal Championships (Highest in 2021)- Pioneering Triple Localization SLAM system for robot movement- Contribution to Open Source Library Road Runner- Custom part manufacturing using 3D Printing

Roles and Accomplishment- Founding and leading Serendipity, Community Outreach Initiative- Started Distance Learning Coding Classes for General and Special Education Students during COVID-19- CAD Designer of Autonomous Robotic Intake- Lead Outreach Visionary and Documenter- Creating a Special Needs Outreach Program specially tailored modularized curriculum- Drive Team, Driver- Created interactive student guides

- Front-End Website Development with Vue JS and Vuetify framework- Back-End Website Development and databases with Firebase by Google

- Developed an Autonomous Germ Detecting Robot- 3D Printed all custom robotic pieces- Programmed the robots Smart Autonomous Movement System- Assembled and configured Hardware such as Odometry(movement system)