David R. Lyons

Operations Leader · AI & Automation Builder

David R. Lyons

I lead teams and build the systems.

For 14+ years I've led operations and logistics teams — from dispatching trains across 2,000+ miles at Union Pacific to running a multi-million-dollar distribution center as a GM today. Along the way I taught myself to build the tools I wished I had: Power BI dashboards, automations, and AI agents that turn messy operational data into faster, smarter decisions. I lead people and I build systems — and I'm happiest doing both.

View my work Resume GitHub
David R. Lyons
01

Experience & Leadership

American Tire Distributors

Full-time · Feb 2024 – Present · Dallas, TX

General Manager Feb 2024 – Present
  • Direct multi-million-dollar distribution center operations, driving a 54.5% jump in company-wide standings through strategic planning and performance-driven initiatives.
  • Built and lead a 30-person team — owning hiring, mentoring, and data-driven training programs.
  • Cut overtime expense 15% and improved inventory variance through targeted cost-saving measures.
  • Lifted warehouse efficiency 10% with custom Excel tooling and streamlined workflows.
  • Raised route-delivery window success 7.3% via driver training, manifest analysis, and DOT compliance.
  • Stood up Power BI dashboards for daily order monitoring and rapid response to disruptions.

Union Pacific Railroad

Full-time · 12 yrs 3 mos · 2011 – 2023

Manager of Yard Operations Nov 2022 – Jun 2023
  • Coordinated 700+ freight cars daily, cutting terminal dwell and improving departure reliability.
  • Reduced safety incidents through field training exercises and Federal Code compliance enforcement.
Corridor Manager Jun 2016 – Nov 2022
  • Led dispatching teams of 10+ across 3 corridors, lifting schedule adherence through a positive culture and strong field support.
  • Supervised transportation plans for Class 1 railroads, reducing downtime several hours per week with better shift planning.
Train Dispatcher Apr 2011 – Jun 2016
  • Directed train traffic across 2,000+ miles of territory, dispatching 15–35 trains daily (Amtrak, intermodal, coal, auto).
  • Enforced safety regulations and intervened in real time to prevent incidents across the network.
02

Selected Work

01

Kansas Castings

AI-driven SEO + cold-outbound engine for an American iron foundry

Problem
A 30-year foundry with thin, unranked web presence and no inbound pipeline — invisible for the searches its buyers actually run.
Built
Rebuilt the site's technical SEO and content end-to-end (titles, schema, copy across every page), shipped a custom RFQ lead-capture form with a full JSON-LD schema stack, then stood up a cold-outbound system: dedicated sending domain with DKIM/SPF/DMARC, warmup, Apollo prospecting, and per-vertical email sequences.
Result
Surfaced in Google's AI Overview for a high-intent buyer query within a week of launch, and the SEO work produced a real inbound RFQ — direct revenue, not vanity metrics.
  • Google AI Overview ranking
  • < 1 week time to rank
  • 100% inbox outbound deliverability
  • real inbound RFQ pipeline
  • WordPress REST
  • Python automation
  • JSON-LD schema
  • Apollo + Smartlead
  • Email deliverability
02

AITrader

Multi-platform autonomous trading agents with eval-first discipline

Problem
Building trading automation that is honest about edge — most bots ship on backtest hopium and blow up live.
Built
Agents spanning five brokerages/markets behind a unified pipeline, an eval-first research harness that validates strategies out-of-sample before any capital moves, hard risk gates (stop-loss, position caps, momentum gates), and scheduled execution via OS task scheduling rather than fragile in-app crons.
Result
A disciplined research-to-execution loop where strategies must clear walk-forward evaluation to run — and unproven ones stay disabled. The engineering rigor is the point.
  • 5 brokerages markets
  • eval-first harness method
  • risk-gated execution safety
  • Python
  • Broker APIs
  • Backtest/eval MCP
  • Task Scheduler
  • Walk-forward eval
03

OpenClaw Agent Stack

Self-hosted multi-agent orchestration across VPS + local

Problem
Running a fleet of always-on AI agents reliably, cheaply, and on infrastructure I control.
Built
Seven specialized agents with distinct workspaces and permissions, hybrid model routing (local GPU + cloud models with fallback), cron-driven automation, and a hardened VPS deployment running as a systemd service alongside a local node — with delivery routed through Telegram.
Result
An owned, resilient agent platform that runs scheduled work unattended and routes each task to the right model at the right cost.
  • 7 specialized agents
  • hybrid local + cloud routing
  • self-hosted VPS deployment
  • Node.js
  • systemd
  • MCP servers
  • Model routing
  • Cron automation
04

Agentic Dev Tooling

A personal AI-development workflow system

Problem
Default AI coding tools waste tokens and lack durable context, repeatable process, and guardrails.
Built
A layered workflow system: custom skills, rules, and hooks that enforce process; a Rust token-optimizer proxy that compresses dev operations; and a persistent file-based memory backend that carries context across sessions.
Result
A faster, cheaper, more consistent agentic engineering loop — with large token savings on routine operations and context that survives between sessions.
  • 60–90% on dev ops token savings
  • skills + rules + hooks process
  • persistent memory context
  • Claude Code
  • Rust
  • Hooks
  • MCP
  • Skill authoring
03

Capabilities

Leadership

  • Team building & hiring
  • Coaching & mentoring
  • P&L / cost management
  • Operations & logistics
  • Safety & compliance

AI / Agents

  • Multi-agent orchestration
  • MCP server integration
  • Model routing & fallback
  • Eval & backtest harnesses
  • Prompt & workflow design

Automation

  • Scheduled job pipelines
  • Browser automation
  • Cron & systemd services
  • Hooks & process guardrails
  • Notification routing

Web / Engineering

  • Astro & static sites
  • React / Electron apps
  • REST API integration
  • WordPress engineering
  • Technical SEO & schema

Data / Integrations

  • Broker & market APIs
  • Google Workspace APIs
  • Email deliverability infra
  • Python data tooling
  • Persistent memory stores
04

Recommendations

“I had the extreme pleasure of working side by side with David as a Joint Corridor Manager at the SDC in Texas. He is far and away the most dedicated individual I have met in the ten years of my railroad career. David has a quick intellect, excellent communication and written skills, and possesses extreme integrity. His contributions were absolutely invaluable, and it seems there's no challenge too big for David to tackle.”
Deidra Winzer Director of Operations Worked with David on the same team
“I worked side-by-side with David in a management role for nearly 10 years. David always goes above and beyond, working outside the box, to resolve issues and save cost, all while producing top-notch results for employees and customers. I highly recommend him for the logistics/operations management fields if your company is looking for new ways to produce an outstanding product — benefiting both customers and employees while saving money.”
Larry Bell Regional Airline Captain Worked with David in a management role for ~10 years
“As a manager at Union Pacific, David exhibited the utmost professionalism, a work ethic second to none, an insatiable desire to learn and improve, strong leadership skills, and an ability to inspire others through empowerment and personal example. David will be a very valuable asset to any organization.”
Matt Lyons Retired from Union Pacific Railroad Was senior to David
“I worked with David for several years on a territory of railroad jointly owned by BNSF and UPRR. I found him to be a very professional, intelligent, diligent, and honorable person who was a great partner. He will excel in any position requiring planning and decision making, and is an excellent motivator of the people around him.”
Jim Hays Manager at BNSF Railway Worked with David on the same team
05

About

I lead operations for a living. Today I run a multi-million-dollar distribution center as a General Manager; before that I spent 12 years at Union Pacific Railroad, working up from train dispatcher to corridor and yard-operations manager — coordinating crews, traffic, and safety across thousands of miles.

Somewhere along the way I got tired of waiting on reports and started building my own — Power BI dashboards, Excel tooling, then automations and AI agents that turn operational data into decisions. What began as scratching my own itch became a second skill set: self-hosted multi-agent systems, eval-gated trading automation, and a growth engine that put a 30-year-old foundry into Google's AI Overview and produced real inbound revenue.

I lead people and I build systems. I care about outcomes over demos, culture as much as numbers, and turning "what if" into "running in production." If that mix is useful to your team, let's talk.