Download PDF ↓
Builder who ships working AI systems end to end. Designed, built, and operates a self-hosted agent platform on a multi-node Linux fleet: an MCP server exposing 21 typed tools, quota-aware model routing across three cloud providers plus a local open-weight model on owned GPU hardware, tiered retrieval over a private corpus, and an evidence contract requiring verifiable output before a job is accepted. 12+ years across IT support, hardware lifecycle, and inventory and logistics operations; current work is high-volume enterprise hardware configuration under production SLAs. AAS in Cloud Computing. Seeking applied AI engineering roles building agents, tool calling, retrieval, and the guardrails around them.
LLM application development; agent orchestration and tool calling (Model Context Protocol); retrieval-augmented generation and semantic search; model routing and cost control; local open-weight inference; guardrails, evaluation gates, and observability; Python back-end and API design; SQLite and schema design; Linux administration; Docker and systemd services; DNS, DHCP, TCP/IP, subnetting, NAT; identity and access control; CI/CD and release discipline; technical documentation and audit trails; incident triage and root cause analysis.
Tower - Agent Orchestration Platform
- MCP server exposing 21 typed tools that turn a multi-node Linux fleet into a pool of AI workers, consumable by any Model Context Protocol client without a second API to maintain.
- Quota-aware model routing: scores three cloud providers on live usage and reserves the winner atomically in one transaction, or routes fully-specified work to a local open-weight model at zero marginal cost.
- Concurrency by resource, not host: mutating jobs declare what they touch and serialize on that claim, so unrelated work runs in parallel while multi-host jobs stay protected.
- Evidence contract: single-use tokens, immutable request and durable response per run, and independent verification, because a zero exit code means the process ended, not that the work happened. Python, FastMCP, SQLite; 308 tests.
Nexus - Agent Observability & Knowledge Base
- Operator surface for the platform: live node health, running and completed agent jobs with the evidence each returned, and Web Push alerting on failure.
- Tiered retrieval over a private corpus: compact summary cards first, a separate vector index over normalized working history second, raw sources only for audit, so lookups stay cheap before they go deep. Python, FastAPI, vanilla JS.
Loupe - Self-Hosted ML Photo Culling
- Full-stack Python/SQLite web application with its own machine-learning pipeline: face clustering, scene and aesthetic scoring across ~80k frames, running on owned GPU hardware via ONNX.
- Network-aware access control and a commit-before-deploy release discipline via systemd.
Self-Hosted Linux Fleet, Multi-Node Infrastructure
Ongoing
- Designed and operate a multi-node fleet (Ubuntu, Synology NAS, Raspberry Pi) across two subnets including a dedicated storage subnet behind a NAT gateway; administer ~11 TB RAID5 storage.
- Docker and systemd services with a high-availability failover design published through Cloudflare Tunnels and Access; failover validated under a real control-plane outage.
Configuration Technician
Sep. 2025 - Present
Future Tech Enterprise, Inc.
Plano, TX
- Global IT solutions partner specializing in high-volume PCaaS configuration, logistics, and lifecycle management for Fortune 500 and government clients.
- Hardware Install & Configuration: install, configure, and maintain enterprise endpoint hardware at volume, covering imaging, driver configuration, and provisioning against per-client build specifications.
- Firmware & Platform Security: configure BIOS/UEFI settings, TPM enablement, Secure Boot, and firmware asset tag values, applied both manually and programmatically.
- Deployment & Decommission: execute the full hardware lifecycle from receiving through decommission and liquidation, including secure data sanitization and disposition of retired assets.
- Quality Assurance & Failure Triage: white glove inspection, functional testing, DOA identification, and root cause analysis prior to release, working to production SLAs.
- Asset Management & Documentation: scan and log serial and service tags into the ERP/inventory system, keeping physical inventory in agreement with system records.
Courier
Oct. 2023 - Sep. 2025
FedEx Express
Sherman, TX
- Sustained physically demanding, time-bound work; daily vehicle and safety inspections with defects reported through required documentation channels.
- Planned and adjusted routes to hold delivery commitments while reducing time and fuel cost.
Driver
Aug. 2020 - Sep. 2023
Uber
Dallas-Fort Worth Metroplex, TX
- Provided safe, reliable transportation with high customer-satisfaction ratings; developed communication and problem-solving skills through constant client interaction.
IT Specialist / Inventory Control Manager
Aug. 2012 - Jun. 2020
Jet Specialty Inc.
Odessa, TX
- Owned inventory control for a high-volume parts and equipment operation at an industrial supply company serving the oil and gas industry.
- Ran regular audits, identified discrepancies between physical stock and system records, and developed corrective actions; used data analytics to forecast needs and reduce excess stock.
- Executed IT support tasks across hardware and software systems to maintain operational efficiency.
Associate of Applied Science (AAS), Cloud Computing, Infrastructure
Collin College · McKinney, TX
Graduated May 2026
- Networking coursework aligned to CompTIA Network+ objectives; Linux installation and configuration; cloud networking and storage; databases; security; scripting with Python; containerization and orchestration (Docker, Kubernetes).
Certifications, Skills & Technologies
Certifications
Cloud Infrastructure Technician (Collin College)HP Commercial & Consumer Desktops, Workstations & Notebooks Service Qualification (SQ24-PC)CS50P: Introduction to Programming with Python (Harvard/edX)Claude 101 (Anthropic)
AI & Machine Learning
Model Context Protocol (MCP)LLM tool callingAgent orchestrationRetrieval-augmented generationSemantic search & embeddingsModel routing & quota managementGuardrails & evaluation gatesContext designLocal inference (Ollama)ONNXClaude CodeCodex
Engineering
PythonFastAPISQLiteBashGitCI/CDREST and tool APIssystemdDockerPodmanLinux (Ubuntu)DNSDHCPTCP/IPSubnettingNATTailscaleCloudflare (Tunnels, Access, Pages)OAuth/JWT
Endpoint & Platform
Windows 10/11Windows ServerActive DirectoryMicrosoft IntuneWindows AutopilotSCCM/MDTPXE BootBIOS/UEFITPMSecure BootVMware/Hyper-V