Cole Wisniewski let's talk
available Spring 2027+ RIT CIT · 2028

DevOps · platform · SRE

I build platforms
that recover.

I turn reviewed commits into observable, self-correcting systems. My work spans Kubernetes, infrastructure as code, delivery automation, and the software that runs on top.

explore the work open GitHub
  • k3s
  • Argo CD
  • OpenTofu
  • Cilium
  • Prometheus
  • Linux
Cole Wisniewski wearing glasses and a dark suit jacket

Rochester, NYavailable Spring 2027+

about Cole

Linux automation.
Container infrastructure.

Cole Wisniewski is a Computing and Information Technologies student at RIT who builds reliable delivery paths from declared infrastructure through application behavior.

Education
RIT · B.S. Computing and Information Technologies · GPA 3.6 · 2024 to present
Experience
Fredonia Central School · Summer IT Support · 2024 and 2025
Availability
Spring 2027 or later
Location
Rochester, New York
  • ship through Gitreviewed, repeatable delivery
  • observe the resultsignals that reach the operator
  • design recovery firstbackups, snapshots, documented tradeoffs
email GitHub LinkedIn

production platform

Small footprint.
Clear boundaries.

One physical host, three isolated guests, and deliberate placement based on latency, exposure, and failure mode.

physical / 01

colewiz

Proxmox VE

  • Ryzen 5 3600X · 6c/12t
  • 32 GB DDR4
  • 8 TB ZFS + 1 TB NVMe
vm / 100

k3s-01

platform node

  • Cilium eBPF dataplane
  • 27 app + 6 infrastructure directories
  • GPU requests time-sliced
lxc / 201

amp

game server host

  • separate network boundary
  • dedicated resource limits
  • outside the web tunnel
lxc / 202

forge

source control

  • Forgejo + LFS
  • LAN only, on purpose
  • repositories on NVMe

platform stack

  • Kubernetes / k3s
  • Argo CD
  • Cilium / eBPF
  • OpenTofu
  • Prometheus + Grafana
  • SOPS + age
  • Cloudflare
  • Proxmox

Placement follows the hardware. Latency-sensitive workloads stay on the 1 TB NVMe. Bulk data goes to the single-disk 8 TB SMR pool, which favors capacity over disk redundancy. Sanoid provides local rollback; PostgreSQL backups leave the host for R2.

selected work

Systems you
can inspect.

Each case study pairs the system with the engineering choice that made it more reliable, maintainable, or useful.

platform engineering

homelab-gitops

Live system · source private by design · verified September 2026

Challenge
Run application and cluster infrastructure on one small platform without making manual state authoritative.
Decision
Use two directory-driven ApplicationSets, ordered sync waves, automated pruning, and self-heal around reviewed commits.
Outcome
One source of truth covers delivery, exposure, observability, policy, and recovery across 33 source directories.
  • Argo CD
  • Kustomize
  • Helm
  • Cilium

33source directories

3mreconciliation

15dmetric retention

platform-topology.svgarchitecture / current
Layered platform topology showing edge, hypervisor, GitOps platform, observability, recovery, and declared source boundaries
Delivery, exposure, and recovery paths shown as one system.
tigerhub / diningdesktop
TigerHub desktop dining view with campus locations, menus, search, and navigation
Dining, menus, and campus context stay in one offline-first flow.
reliabilitysource ↗

tigerhub

Challenge
A hosted collection dependency introduced an unreliable production failure mode for a campus app.
Decision
Move collection on-device and make the Flutter application offline-first.
Outcome
The hosting dependency left the critical path, with application behavior backed by 445 tests.
  • Flutter
  • Dart
  • offline-first
  • 445 tests
wizsearch / resultsdesktop
WizSearch desktop results view with source filters, media previews, attribution, and collection controls
Source context, attribution, previews, and collection controls share one view.
desktop productsource ↗

wizsearch

Challenge
Multi-source results lose value when source context, attribution, and collection controls are separated.
Decision
Keep the workflow local-first in a Rust and Tauri 2 desktop interface.
Outcome
Filters, previews, attribution, and collection controls remain together in the results view.
  • Rust
  • Tauri 2
  • local-first
infrastructure as code

infra-tofu

Cloudflare, Proxmox, and GitHub resources declared with remote state and encrypted variables.

review the IaC ↗

operating practice

Delivery is only
the first step.

The platform is treated as an operating system. Every change needs a delivery path, a useful signal, and a recovery path.

deliver

3mreconciliation

Git stays authoritative.

Two directory generators turn application and infrastructure folders into Argo CD Applications. Automated pruning and self-heal return the cluster to its declared state.

  • ApplicationSets
  • self-heal
  • Kustomize

observe

15dmetric retention

Alerts leave the dashboard.

Prometheus retains cluster and workload signals under an 8 GB cap. Alertmanager routes actionable alerts to ntfy, so failures reach a phone.

  • Prometheus
  • Alertmanager
  • Grafana
  • ntfy

recover

30ddatabase retention

State has an exit plan.

Five PostgreSQL workloads archive WAL and daily base backups to R2. Sanoid snapshots cover file data so database and storage recovery meet at a known point.

  • CloudNativePG
  • R2
  • Sanoid
  • ZFS

architecture decisions

The tradeoff belongs
in the portfolio.

Seven records document context, choice, consequences, and the condition that would justify reversal.

read all 7 records
004

networking

Cilium replaces three components

One eBPF dataplane replaces overlapping network layers.

tradeoff

More capability, more responsibility in one component.

007

service mesh

No mesh on one node

The extra latency and policy layer do not repay their cost.

revisit when

The cluster gains nodes or service identity becomes required.

resume

The one-page version.

A concise record of the systems, software, and operational choices shown throughout this portfolio.

Education
RIT · B.S. Computing and Information Technologies · GPA 3.6 · expected 2028
Focus
DevOps · platform engineering · site reliability
Availability
Spring 2027 or later
download resume open PDF

The embedded preview may not appear in every mobile browser. The complete one-page PDF is available through either link.