Darrell S. Best Jr.

Read My Articles Contact Me

01. Professional Summary

  • AI Research Engineer with over a decade of programming experience and 7+ years delivering complex AI/ML solutions.
  • Led AI projects including multilingual LLMs, federated learning systems, FPGA bitstream error correction, and medical text classification.
  • Proven track record in using PyTorch, Hugging Face, DeepSpeed, and Flower to deliver scalable, secure, and efficient AI systems.
  • Developed mission-critical AI solutions for defense, healthcare, and telecommunications, driving operational efficiency and innovation.

02. Work Experience

USC ISI · Featured Project 2019 – Present

Hawkeye

Built a scalable multilingual chat system supporting 4+ foreign languages, enabling real-time cross-language communication for defense applications.

GPT-2 DeepSpeed PyTorch Hugging Face NLP
USC ISI

Danube

Engineered federated learning pipelines that train models across distributed nodes without centralizing sensitive data, with built-in anomaly resistance.

Flower PyTorch Federated Learning
USC ISI

Sonic Screwdriver

Created a novel FPGA bitstream error correction system using custom tokenization and masked language modeling, applying NLP techniques to hardware repair.

Hugging Face MLM FPGA
USC ISI

HealthMap

Automated mapping of unstructured medical reports to ICD diagnostic codes, accelerating clinical workflow and improving classification accuracy.

NLP Healthcare AI Text Classification
USC ISI

GreenSight

Engineered text encoders that classify content by Schwartz’s 19 universal moral values, enabling nuanced analysis of cultural narratives at scale.

Hugging Face Text Encoding Values AI
QinetiQ US 2018 – 2022

HMDS & IGSR

Built IED detection systems for Army Husky vehicles using ground-penetrating radar and developed border-crossing detection for the FBI using ResNet CNNs.

Computer Vision ResNet Defense

Junior Software Engineer, Windstream — Greenville, SC

Sep 2017 – Mar 2018

Built end-to-end provisioning software (PUMA) for DSLAMs and network devices using multiple databases and remote connections.

Research Assistant, Clemson University — Clemson, SC

May 2014 – Feb 2018

Solved the Midas touch problem in eye-tracking with natural eye gestures. Published A Rotary Dial for Gaze-based PIN Entry at ETRA 2016.

03. Education

MS in Computer Science

2024 – Present
  • University of Southern California, Viterbi School of Engineering
  • Focus: Data Science

Graduate Coursework

  • CSCI 561 — Foundations of Artificial Intelligence: Search algorithms, constraint satisfaction, probabilistic reasoning, and game-playing agents.
  • CSCI 570 — Analysis of Algorithms: Advanced algorithm design and complexity analysis including dynamic programming, graph algorithms, and NP-completeness.
  • CSCI 544 — Applied Natural Language Processing: Neural language models, sequence-to-sequence architectures, transformers, and text classification pipelines.
  • CSCI 585 — Database Management Systems: Relational and NoSQL databases, query optimization, transaction processing, and distributed data systems.
  • CSCI 567 — Machine Learning: Supervised and unsupervised learning, kernel methods, ensemble techniques, and deep learning fundamentals.

04. Publications

A Rotary Dial for Gaze-based PIN Entry

Best, Darrell S. and Duchowski, Andrew T. (2016). In Proceedings of the Ninth Biennial ACM Symposium on Eye Tracking Research & Applications (ETRA '16), pages 69–76. ACM.

https://doi.org/10.1145/2857491.2857527

View on Google Scholar

05. Skills

Languages

Python JavaScript C++ SQL Bash

AI / ML

PyTorch Hugging Face TensorFlow DeepSpeed Flower scikit-learn

DevOps & Tools

Docker Git CI/CD Linux

06. Contact

Open to collaborations in AI research and engineering. Feel free to reach out.

Email:

LinkedIn: linkedin.com/in/darrellsbest

Scholar: Google Scholar

GitHub: github.com/DarrellBest