Zac Kelder · Systems & infrastructure engineer

Curiosity is a serious thing.

I’ve spent eight years operating infrastructure where reliability matters, from deployable military systems, networks and communications to enterprise Linux, Windows, VMware and identity services. I bring that operational perspective to the platforms I build, spanning cloud infrastructure, automation, computer vision and quantitative research.

Open to systems administration · infrastructure engineering · cloud operations · identity and access

See the work Download résumé (PDF) Professional experience
Wildlife observation platform · Edge systems

MockingBird

A local-first wildlife observation platform that turns cameras and footage into a searchable record of a place.

MockingBird ingests live and recorded video, detects and tracks animals, and organizes the activity into Visits, a Bird Log and a daily view of what happened. It runs on your own computer with local sign-in and no cloud account or service. I designed and built the whole stack: a React interface, a FastAPI backend, an ONNX computer-vision runtime, SQLite persistence and a Raspberry Pi Camera Node.

Read the case study
MockingBird application showing a Visit page: a zoomable evidence frame of a nestling in green foliage, a banner noting three other subjects detected in the same encounter, and a suggestion panel beside the sidebar navigation for Today, Cameras, Review Queue, Bird Log, Place and Settings.
A Visit in the running application: the best evidence frame, tracked subjects from one shared encounter, and an identification suggestion. Real product on a demo instance fed public-domain clips.

  • Camera and edge systems

    A Raspberry Pi Camera Node, RTSP/ONVIF and USB cameras, still-image URLs and imported footage, with Zones to focus each view and supervised capture that recovers from failures.

  • Wildlife intelligence

    MegaDetector v5a on ONNX Runtime and a tracker assemble live or recorded frames into Visits with their best evidence. Optional BioCLIP 2 and BirdNET suggestions stay labelled as evidence, and a person confirms what enters the record.

  • Application and local operations

    Today, live views, Visits, an Activity Library, the Bird Log and Insights, on SQLite you control with local sign-in, health alerts, retention, and verified backup and restore. No cloud account or service, and no telemetry.

Pre-release · v0.1.0. Validated on a Windows 10 host with an NVIDIA GPU and a Raspberry Pi 4 Camera Node.

Stack
FastAPI · SQLite · React · ONNX
Edge
Raspberry Pi Camera Node

Quantitative research platform · Data and systems engineering

Apex

A quantitative research platform that collects real market data, evaluates strategies point-in-time, and records every decision.

I designed and built the whole system. Scheduled workers collect public crypto, prediction-market, SEC and options-chain data into durable PostgreSQL evidence. Strategies are evaluated against only what was known at the time, and qualifying ones run in PAPER trials with risk controls, double-entry accounting and an Operator Console. Research and simulated capital only; the public demo is synthetic.

Key engineering Database-enforced evidence integrity, unattended systemd operation with startup safeguards and recovery, and one engineer owning everything from the timers to the React console.

Read the case study Open the interactive demo
  1. 01EvidenceOnly what was known before the decision cutoff.
  2. 02ProposalA strategy proposes; an allocator sizes it.
  3. 03Risk kernelFixed rules accept or reject, with no bypass.
  4. 04LedgerExact money math in an append-only record.
Apex demo outcome cards, including a settled trade, a rejection by allocation policy and a rejection by the risk kernel, which is selected.
Outcomes, rejections included. Real output from the static synthetic-data demo, where each run is reproducible and nothing is live.

Cloud infrastructure · AWS architecture

ArenaOps

An AWS backend that wakes on demand, reports readiness to the visitor, runs its workload and returns to idle. ECS/Fargate sat at desired count zero between sessions, with Terraform-provisioned infrastructure, IAM/OIDC authentication and GitHub Actions delivery through ECR to ECS. Retired in August 2026 to end the load-balancer cost; architecture, screenshots and cost record preserved.

Key engineering A scale-to-zero lifecycle with visible readiness, and a reviewable Terraform and CI/CD path from commit to running container.

ArenaOps connected multiplayer room with operator controls and the active arena view.
Realtime WebSocket room after health checks pass.
Read the case study Architecture and cost record

iOS application · Backend and cloud operations

Food Checker

An iOS app that scans a food label and flags the ingredients you choose to watch for, delivered as a complete system: Expo / React Native client, OCR, configurable rules, a hardened FastAPI service with Supabase authentication and data, Terraform-provisioned AWS, automated releases and private Prometheus and Grafana observability. In TestFlight beta.

Key engineering Credential-light releases through GitHub Actions, OIDC and SSM, verified by a health check and a real analysis request after every deploy.

Food Checker iOS scan screen with ingredient label guidance and scan controls.
Scan
Food Checker iOS history screen showing structured scan results and risk states.
Review
Review the engineering TestFlight beta

Eight years in accountable environments.

The professional record: military field systems and a team to lead, then Linux, VMware and Microsoft enterprise infrastructure.

Technical operationsEnterprise and U.S. Army environments
8+ yrs
Employees supportedGlobal workforce, escalations and after-hours incidents
≈400
EndpointsProvisioned and tracked across multiple sites
≈300
Equipment accountabilityLed a nine-person technical team
$20M+

Certifications

  • CompTIA Security+
  • CompTIA A+
  • ITIL Foundation
Download résumé (PDF)
2023—2025Trilliant Networks · Oct 2023 – May 2025

Systems Administrator

Linux, VMware and Microsoft enterprise infrastructure, identity and audit remediation.

  • Linux and VMware. Administered Linux development systems (RHEL, Rocky Linux and Ubuntu) in my assigned environment: patching, service and log troubleshooting, sudo and group permissions, backup and recovery. Deployed Linux and Windows server VMs through vCenter and worked through snapshots, migrations and ESXi problems. Scripted repeatable work in Bash and PowerShell.
  • Identity and access. Administered Red Hat IdM / FreeIPA, with LDAP, Kerberos and SSSD, for centralized Linux authentication. Administered Microsoft 365, Active Directory and Entra ID, including SaaS single sign-on, MFA, licensing and access reviews.
  • Endpoints. Provisioned and tracked the multi-site endpoint fleet, and supported Intune and Autopilot rollouts, onboarding and offboarding, and application and OS updates.
  • Email security and credentials. Investigated email threats and security alerts, tuned Exchange Online and Defender filtering, and supported Purview data-loss prevention. Implemented Keeper SSO and privileged-credential protection for elevated accounts and for systems that could not use centralized identity management.
  • Operations. Supported infrastructure, identity services and end users across the global workforce: escalations and after-hours incidents under formal SLAs, plus coordination with infrastructure teams, vendors and stakeholders on server upgrades, server-room changes and recovery.
  • Security audits. Gathered screenshots and technical evidence for security audits and demonstrated configured controls to auditors. Drafted IT and security policies and control documents for supervisory review. Personally implemented the technical changes needed to remediate findings.
2017—2023U.S. Army · Jan 2017 – Oct 2023

Systems Maintainer / Technical Team Lead

Deployable field infrastructure, leadership and operational readiness.

  • Leadership. Led the technical team responsible for the readiness, maintenance and troubleshooting of mission-critical IT and communications equipment.
  • Field infrastructure. Deployed and integrated field IT environments end to end, from generator power and distribution through bare-metal Windows and Linux servers, Cisco networking, SATCOM links, operational systems and user laptops.
  • Build and readiness. Built, imaged, configured and tested servers, networks and workstations for exercises and deployments, and coordinated preventive maintenance, inspections and readiness checks.
  • Fault isolation and restoration. Diagnosed and restored faults across interconnected power, server, network, satellite and endpoint systems during time-sensitive field operations.
  • Training and communication. Trained maintainers, documented troubleshooting procedures, coordinated equipment upgrades, and briefed users and leadership on status and restoration needs.

An operator’s perspective on engineering.

Eight years of running systems that people depend on, from U.S. Army field infrastructure to enterprise Linux, VMware, Microsoft and identity services, shaped how I engineer. I think first about operational judgment: what happens when a dependency fails, how a system is recovered, who is allowed to change it, and whether the next administrator can maintain it.

That perspective runs through the systems I design and build end to end: AWS infrastructure delivered with Terraform and CI/CD, a research platform that fails closed, a computer-vision application with an edge camera node, and a mobile app with its own API and monitoring. Each one owns its security boundaries, recovery paths and documentation. Where a repository is private, its case study carries the architecture and evidence instead.

  1. Check health before declaring success.
  2. Make failure states legible to users and operators.
  3. Prefer controlled, repeatable delivery paths.
  4. Document boundaries, recovery, and limitations.

Bring a curious engineer to your next system.

Systems administration, infrastructure engineering, cloud and platform operations, identity and access, security operations, and technical support engineering roles. Email is the quickest way to reach me, and I’m glad to send a résumé or walk through any project.

(845) 802‑6510 linkedin.com/in/zkelder github.com/zkelder

Download résumé (PDF) Two pages, October 2026.