Leoul Meaza

Currently

AI Trainer

Handshake

Oct 2025 - Present · Remote

Selected for the Handshake AI Fellowship to develop and evaluate domain-specific prompts to assess the performance of large language models (LLMs), and analyze LLM outputs for scientific accuracy, clarity, and depth in specialized subfields, and contribute to improving AI understanding of various topics through expert review and feedback, and conduct independent research to support prompt development and evaluation tasks.

Recent Projects

Bontoo

Dec 2025 - Jan 2026

Architected a secure serverless backend for a custom AI chat application, using Node.js to proxy Anthropic API requests and decouple secret management from the client, automated production deployments by establishing a CI/CD pipeline via GitHub, and ensured application resilience by engineering a backend to validate JSON payloads.

Node.jsVercelJavaScriptAnthropic APIGitHTMLCSS

View Project

Today I Learned

Jul 2025 - Aug 2025

Built a full-stack web app to submit, categorize, and vote on bite-sized learning facts. Utilized Supabase real-time functions to support live database updates, voting mechanisms, URL validation, and category filtering.

ReactSupabaseJavaScriptCSSNetlify

View Project

Real-Time Microplastic Surge Predictor

Jan 2025 - May 2025

Architected a novel AI-powered AWS workload that ingests live microplastic sensor data and rainfall forecasts into Amazon SageMaker, predicting contamination surges before they occur. Presented the research project to faculty, securing an Elevate Student Research Mentor role.

AWSMachine Learning

Experience

Student Research Mentor

Radford University — Elevate Research Program

Aug 2025 - Dec 2025 · Radford, Virginia

Guided 30 students in the Foundations of Cloud Computing class through an AWS cloud computing research project. Held office hours and responded promptly to emails to troubleshoot issues, clear roadblocks, and help students deliver higher-quality work.

Software Engineer - AI Consultant

Outlier (Scale AI)

Feb 2024 - May 2024 · Remote

Debugged and validated code produced by large language models, identifying and resolving issues to ensure reliable outputs. Leveraged expertise in Java, JavaScript, and Python to refine and adapt LLM-generated code, improving usability and compatibility across various platforms.