Portfolio

Owning Your Agent Part 2 — local agent stack with Hermes on a single RTX 3090, pushing consumer hardware to its limits
  • Hermes
  • Single GPU
  • Agent

Owning Your Agent Part 2 — Local Agent with Hermes on a Single RTX 3090; pushing the absolute limits
Local LLM Inference Lab — reproducible vLLM, llama.cpp, multimodal, and agentic workload research on dual RTX 3090s
  • vLLM
  • llama.cpp
  • Benchmarks

Local LLM Inference Lab — Multi-Model Benchmarking with vLLM, llama.cpp & Agentic Workloads
Local LLM Infrastructure V2 — dual RTX 3090 Threadripper Pro AI server with 128 GB RAM
  • Infrastructure
  • Server
  • LLM

Local LLM Infrastructure V2 — Dual RTX 3090 Threadripper Pro AI Server
Owning Your Agent Part 1 — local agent stack with Hermes Agent + Qwen3.5-27B + GLM-4.7-Flash on dual RTX 3090s
  • Hermes
  • Local AI
  • Agent

Owning Your Agent Part 1 — Local Agent with Hermes on Dual RTX 3090s
Dual RTX 3090 local LLM infrastructure build with 48 GB VRAM
  • Infrastructure
  • CUDA
  • LLM

Local LLM Infrastructure — Dual RTX 3090 Cluster, 48 GB VRAM
Melange — Apple Silicon LLM memory profiler and KV cache analyzer built in Rust
  • Rust
  • Apple Silicon
  • LLM

Melange — Apple Silicon LLM Memory Profiler & KV Cache Analyzer (Rust)
Personal portfolio website project — HTML, CSS, and JavaScript built by Zachary Cangemi
  • HTML
  • CSS
  • JS

Personal Portfolio Website