// systems · programs · agents

Ryan Anstett

Systems Administrator · Program Manager · AI Developer @ Trellix

I build enterprise AI agents, lead large-scale platform migrations, and ship data-driven side projects.

See the work ↓ github linkedin
Selected Work · graded nightly

Models that grade themselves.

MLB Betting Model
mlb @ mlb-algo
A daily MLB model board. Prices every game from season run environments and probable-starter FIP, then shops the moneyline and total across DraftKings and FanDuel to surface Kelly-sized edges — with a hard edge gate, the model's blind spots listed next to its picks, line-movement charts from its own odds archive, and paper-tracked bet and model records graded nightly against final scores.
PythonNext.js MLB Stats APIDraftKings / FanDuel oddsSupabase
↗ live board
units · season
graded nightly
College Football Betting Model
cfb @ cfb-algo
A weekly college football model. Composites ~15 public prediction systems with SP+ and FPI power ratings, prices every FBS game against FanDuel spreads, totals, and moneylines, and posts a weekly capper card — unit-sized plays plus lock and moonshot parlays — while archiving every line for closing-line-value grading.
PythonNext.js ESPN APIFanDuel oddsSupabase
↗ live board
units · season
graded vs close weekly
NFL Betting Model
nfl @ nfl-algo
A weekly NFL model. Blends ~40 public prediction systems with ESPN FPI and EPA unit ratings, prices all 16 games against FanDuel and DraftKings spreads and moneylines, and posts a weekly capper card — unit-sized plays plus lock and moonshot parlays. Every game gets its own written report page: the setup, the power ratings, the computer consensus, the market read, and the model's verdict.
PythonNext.js ESPN APIFanDuel oddsSupabase
↗ live board
units · season
graded vs close weekly
Golf Betting Model
glf @ golf-algo
A PGA Tour betting-value engine. Pulls the field and live book odds, then blends season strokes-gained, recent form, and course fit into one skill rating, Monte-Carlo simulates each tournament, and de-vigs the market to surface +EV bets and head-to-head edges. Includes an in-tournament live model, form and course-fit backtesting, and a closing-line-value bet tracker.
PythonNumPy Monte CarloESPN / PGA dataSupabase
↗ live board
units · season
CLV bet tracker
0
engineers enabled with AI tooling
0
Jira projects migrated
0
Confluence spaces migrated
0
IT requests automated
About

Enterprise scale, side-project speed.

Enterprise systems administrator and program manager at Trellix with production experience across the Atlassian platform, GitHub Enterprise, and IT service delivery. Recently expanded into AI development — building agents on Google Gemini backed by Jira Service Management that enable conversational IT service requests across the enterprise. I've led platform migrations (on-prem to AWS, cloud to SaaS, Microsoft 365 to Google Workspace) and provisioned 1,000 engineers to GitHub Copilot with live cost reporting to executive leadership.

Experience

Where I've shipped.

Current
Systems Administrator · Program Manager · AI Developer @ Trellix
Promoted to full-time after starting part-time; scope has since expanded to TechOps program management for Engineering plus AI development.
  • Designed and built AI agents on Google Gemini backed by Jira Service Management — enabling conversational IT service requests across the enterprise; replicated with LangGraph + GCP Agent Builder on Cloud Run
  • Provisioned 1,000 engineers to GitHub Copilot with live usage/cost dashboards reported to executive leadership
  • Lead administrator for the Atlassian suite (Jira, Confluence, JSM) across all Trellix Engineering
  • Migrated 400+ Jira projects and 1,600+ Confluence spaces from AWS-hosted cloud to Atlassian SaaS
  • Automated JSM ticket fulfillment (1,000+ requests) and wired SCIM/SAML between Okta and GitHub for automated user lifecycle
  • Program manager for the TechOps practice — presenting Smartsheet dashboards to leaders across Finance, HR, and Customer Success
Previously
Enterprise Workplace Technology (Part-Time) @ Trellix
  • Contributed to the enterprise Microsoft 365 → Google Workspace migration (onboarding, Google Sites, platform config)
  • Supported GitHub repository consolidation following the McAfee Enterprise + FireEye combination
  • Handled DevOps platform requests across Jira, Confluence, GitHub, and Jenkins
Education · B.S. in Engineering — Information Technology · University of Central Florida