Portrait of Davis Byrd

Davis Byrd

AI/ML Engineer

Lake Charles, LA

AI/ML engineer built on 10+ years of hands-on IT systems experience — now applying that troubleshooting and systems design background to Integral AI, where I design and deploy custom AI business systems for small professional practices

About

Starting from the ground up in IT — systems analysis, network infrastructure, and technical support taught me to diagnose problems methodically and keep things running reliably under real-world constraints. I'm now channeling that same discipline into software and AI development, learning hands-on to build AI systems that are just as stable and dependable for the businesses using them.

As Integral AI grows, I'm deepening my knowledge and expertise in ML engineering alongside it.

  • Systems-first thinking

    A decade of IT and network troubleshooting shapes how I debug and design AI systems today.

  • Learn by building

    Every new AI/ML concept gets tested in a real project, not left as a notebook exercise.

  • Honest project status

    Work is labeled by its real stage — prototype or in-progress — not oversold.

Skills

AI & Machine Learning

  • Supervised & Unsupervised Learning
  • Deep Learning
  • Convolutional Neural Networks (CNNs)
  • Ensemble Methods
  • Transfer Learning
  • Model Optimization & Tuning
  • Natural Language Processing (NLP)
  • Large Language Models (LLMs)
  • Prompt Engineering
  • Generative AI

Languages & Libraries

  • Python
  • Java
  • C
  • C++
  • SQL
  • NumPy
  • Pandas
  • Matplotlib
  • Seaborn
  • Plotly
  • Scikit-learn
  • TensorFlow
  • Keras
  • PyTorch
  • OpenCV
  • Ultralytics YOLO
  • NLTK

Tools

  • Git
  • Jupyter Notebook
  • VS Code
  • Kaggle
  • Anaconda
  • Hugging Face

Experience

Systems Analyst / Administrator

May 2015 – Present

Computer-Line, LLC — Lake Charles, LA

  • Delivered IT support for 10+ clients, resolving 15+ weekly requests across Windows/macOS environments, with a focus on efficient troubleshooting and automation of recurring issues
  • Provided end-to-end support for hardware, software, and networking, resolving most client issues within 24 hours by leveraging diagnostic data and performance metrics
  • Administered and maintained 150+ workstations across diverse environments, using RMM tools (LogMeIn) for proactive monitoring, data collection, and predictive maintenance
  • Developed and optimized scripts (PowerShell, Python) to automate system monitoring, data extraction, and reporting, reducing manual workload and improving response time

Projects

Projects are shown at their real stage of progress — working prototype or in-progress build — not as finished products.

RepVision

Working Prototype

A computer vision system that automates exercise classification and repetition counting from video input, eliminating the need for manual tracking and establishing a foundation for future enhancements such as form evaluation.

  • OpenCV
  • YOLOv11 (classification)
  • YOLOv11-Pose (keypoints)
  • Python

An end-to-end exercise analysis pipeline combining OpenCV for video frame extraction and preprocessing, YOLOv11 for exercise classification, and YOLOv11-Pose for keypoint detection and angle calculations. Currently being extended to implement rep counting across all classified exercise types.

View on GitHub ↗

AI Receptionist — Voice Booking Demo

Working Prototype

A browser-based AI voice receptionist that answers scheduling questions, books real appointments, and hands owners a same-tool staff assistant — no telephony required to demo it.

  • FastAPI + WebSocket voice loop
  • Claude Messages API (tool use)
  • SQLite
  • Whisper / Deepgram (STT)
  • Piper / Kokoro / ElevenLabs (TTS)
  • FastMCP staff tools

Answers natural-language availability questions, books an appointment end to end with a spoken confirmation, handles mid-conversation corrections, and exposes the same booking logic to Claude Desktop via MCP so an owner can ask "who hasn't confirmed this week?" as a daily tool.

CPA RAG Assistant — Document-Grounded Q&A

In progress

A retrieval-augmented assistant for a CPA firm's internal documents, built around the part that actually matters for a firm to trust it: guardrails, access control, and citations — not retrieval alone.

  • ChromaDB (embedded vector store)
  • Custom chunking / retrieval pipeline
  • Claude or Groq (togglable LLM)
  • Role-based access filtering
  • pytest

Answers questions grounded in a firm's own documents with citations back to the source, refuses to answer when retrieval confidence is too low instead of guessing, and filters retrieved content by the asking user's role.

Education

Click either image to view it full size.

Bachelor of Science in Applied Computer Science diploma, McNeese State University, awarded to Davis Connolly Byrd

B.S. in Applied Computer Science

McNeese State University — Lake Charles, Louisiana

Conferred May 2023

Relevant coursework: Data Structures & Algorithms, Operating Systems, Database Systems, AI, Computer Networks, Software Engineering

Artificial Intelligence & Machine Learning Bootcamp certificate, Louisiana State University, awarded to Davis Connolly Byrd

AI & Machine Learning Program

Louisiana State University

January 2025 – June 2025

Immersive, active-learning program building proficiency in Python, Keras, and TensorFlow, and a working understanding of machine learning, deep learning, and data science processes.

Get in touch

Open to AI/ML engineering roles and collaborations — feel free to reach out.