
HARRISON OLVERA / AI ENGINEERING OPEN TO REMOTE + DFW HYBRID ROLES
APPLIED AI • LLM PRODUCTS • AGENT WORKFLOWS
I BUILD AIPRODUCTS THATDO REAL WORK.
I connect models to data, tools, and human decisions—with clear limits, useful interfaces, and evidence you can inspect.

SELECTED PROJECTS
View case studiesFOUNDATION
BACKGROUND
& PRACTICE
EDUCATIONBachelor's Degree in Software Development
DeVry University
EXPERIENCEWeb Developer
Vercity Consulting · Power Automate, Power Apps, SharePoint, PowerShell, and JavaScript PDF workflows
RAGstructured outputstool callingMCPmodel routinghuman-in-the-loopTypeScriptNext.jsReactPythonNode.jsREST APIsPostgreSQLpgvector
WORKING METHOD
AGENT DELIVERY
PROCESS
- 01Define
Clarify the user, constraint, and observable outcome.
- 02Ground
Map available data, tools, interfaces, and failure boundaries.
- 03Build
Implement the smallest end-to-end working system.
- 04Evaluate
Test behavior, edge cases, integrations, and evidence.
- 05Document
Make the architecture, limits, and inspection path clear.
SELECTED SCHOOL WORK
Original portfolioFour foundations that still show up in my AI systems work: security, physical interfaces, data pipelines, and testable software.
01SEC285
Security Operations Lab
Linux, Ubuntu, and Kali virtual machines used for MFA, PGP encryption, firewall policy, Nmap scanning, security policy design, and vulnerability assessment.
View original course deck02CEIS101
IoT Smart-Home System
An ESP32 and Arduino system for door-state sensing, intruder-distance monitoring, and automatic lighting, tested with Tinkercad and physical IoT components.
View original course deck03CEIS110
Python Weather Data Pipeline
A Python workflow that downloaded weather data into a database, queried it with SQL, exported and cleaned CSV data, and produced graphical analysis.
View original course deck04CEIS150
Python Stock Application
A structured Python application developed from class diagrams through a text interface, stock records, daily data, summary reporting, and unit tests.
View original course deckWHAT I OPTIMIZE FOR
BUILD THE SYSTEM.
SHOW THE RECEIPTS.
01
Grounded outputRetrieval and citations over unsupported answers.
02
Bounded actionLeast-privilege tools and explicit approval gates.
03
Observable deliveryTests, logs, and release evidence recruiters can inspect.
AVAILABLE FOR THE RIGHT ROLE
BUILD SOMETHING
USEFUL.
Open to conversations about AI engineering, LLM products, automation, and agent systems.
Designed and engineered as an evidence-first portfolio.