Jacob Stevenson

Research software engineer and AI consultant

I translate ideas and expertise into software that works: AI systems, search, websites, and mobile apps for researchers, startups, and the teams around them.

Projects and experience

Melody Search Index

Sept 2023 – now

Research software engineer, BYU music research

  • Solo-built a full-stack melody-similarity search, with OCR and RAG, over a 10k+ record dataset.
  • Cut indexing time 99% (overnight to minutes) and query time 96% (minutes to seconds) for researchers worldwide.
  • Presented the work at conferences in Texas, Salzburg, and Venice.

FMC Research Assistant

2026 – now

Search for 19th-century French music press

  • Built keyword and semantic search (PostgreSQL full-text, bge-m3 embeddings, reranking) over 4,033 documents and 16,574 pages, running on local GPUs.
  • Retired AI-written answers at the scholars’ request; every result is a quoted, cited source page.

AI vs. human sentiment analysis

LLM evaluation for BYU religion researchers

  • Built an LLM pipeline that codes newspaper articles on the researchers’ own schema and measures its agreement with human coders.

Yellowstone field guide app

2026 – now

Offline Android app for a BYU field course

  • Building an offline field guide from a 40-page microbiology guidebook, checked by 1,243 automated content assertions.

Auto-cropper

Computer vision

  • Automated cropping for 7k+ images: 25× faster, with 96% less manual editing for a team of five.

Biosignal research

Signal processing for “Evaluating EEG Monitoring of Standard TMS Intervention for Anxiety”

  • Used FFT and statistical modeling on EEG and heart-rate data to separate state anxiety from trait anxiety.

All projects at jistevenson.com/work

Education

B.S. Computer Science, Brigham Young University

Coursework: advanced software construction, advanced programming, web programming, algorithms, computer systems, software design, systems programming, mathematical cryptography, and computer security.

Talks

  • A Cohesive Look at RAG for Research · Venice, Italy · 2026
  • Multi-faceted Strategies for MEI Corpus Building · IAML Congress · Salzburg, Austria · 2025
  • Simplifying MEI for Searching Applications · Texas, USA · 2024

Skills

Languages
Java, C#, C and C++, Python, JavaScript and TypeScript, SQL
AI and data
LLMs, RAG, Vector embeddings, Semantic search, Elasticsearch, OpenCV, Signal processing
Platforms
AWS, Docker, Linux, Git, PostgreSQL, REST APIs, Android
Exploring
Cryptography, Zero-knowledge proofs, IoT, Neural networks, Integer programming

Also

English and Spanish. Eagle Scout, black belt, musician, DM, philosopher, zymologist, scuba diver, and former “World’s Youngest Person.”