Instructions to use chrisarseno/csuite-model-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use chrisarseno/csuite-model-gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="chrisarseno/csuite-model-gguf",
	filename="csuite-business-32b-Q5_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use chrisarseno/csuite-model-gguf with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf chrisarseno/csuite-model-gguf:Q5_K_M
# Run inference directly in the terminal:
llama cli -hf chrisarseno/csuite-model-gguf:Q5_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf chrisarseno/csuite-model-gguf:Q5_K_M
# Run inference directly in the terminal:
llama cli -hf chrisarseno/csuite-model-gguf:Q5_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf chrisarseno/csuite-model-gguf:Q5_K_M
# Run inference directly in the terminal:
./llama-cli -hf chrisarseno/csuite-model-gguf:Q5_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf chrisarseno/csuite-model-gguf:Q5_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf chrisarseno/csuite-model-gguf:Q5_K_M

Use Docker

docker model run hf.co/chrisarseno/csuite-model-gguf:Q5_K_M

LM Studio
Jan

vLLM

How to use chrisarseno/csuite-model-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "chrisarseno/csuite-model-gguf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "chrisarseno/csuite-model-gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/chrisarseno/csuite-model-gguf:Q5_K_M

Ollama
How to use chrisarseno/csuite-model-gguf with Ollama:
```
ollama run hf.co/chrisarseno/csuite-model-gguf:Q5_K_M
```

Unsloth Studio

How to use chrisarseno/csuite-model-gguf with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for chrisarseno/csuite-model-gguf to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for chrisarseno/csuite-model-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for chrisarseno/csuite-model-gguf to start chatting

How to use chrisarseno/csuite-model-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf chrisarseno/csuite-model-gguf:Q5_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "chrisarseno/csuite-model-gguf:Q5_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use chrisarseno/csuite-model-gguf with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf chrisarseno/csuite-model-gguf:Q5_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default chrisarseno/csuite-model-gguf:Q5_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use chrisarseno/csuite-model-gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf chrisarseno/csuite-model-gguf:Q5_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "chrisarseno/csuite-model-gguf:Q5_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Docker Model Runner
How to use chrisarseno/csuite-model-gguf with Docker Model Runner:
```
docker model run hf.co/chrisarseno/csuite-model-gguf:Q5_K_M
```

Lemonade

How to use chrisarseno/csuite-model-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull chrisarseno/csuite-model-gguf:Q5_K_M

Run and chat with the model

lemonade run user.csuite-model-gguf-Q5_K_M

List all available models

lemonade list

C-Suite Executive AI — GGUF Models

Fine-tuned Qwen 2.5 32B Instruct models for the C-Suite AI executive team. Each model embodies specialized executive personas with distinct codenames, domain expertise, and personality traits.

Models

File	Size	Description	Executives
`csuite-merged-32b-Q5_K_M.gguf`	~22 GB	Base identity + personality (all executives)	All 16
`csuite-technical-32b-Q5_K_M.gguf`	~22 GB	Technical domain specialist	CTO (Forge), CEngO (Foundry), CIO (Sentinel), CSecO (Citadel)
`csuite-business-32b-Q5_K_M.gguf`	~22 GB	Business domain specialist	CFO (Keystone), CRevO (Compass), CSO (Beacon), CPO (Blueprint)
`csuite-operations-32b-Q5_K_M.gguf`	~22 GB	Operations domain specialist	CoS (Overwatch), COO (Nexus), CDO (Index), CRO (Axiom)
`csuite-governance-32b-Q5_K_M.gguf`	~22 GB	Governance domain specialist	CComO (Accord), CRiO (Vector), CCO (Echo), CMO (Aegis)

Quantization

Method: Q5_K_M (5-bit k-quant, medium)
Format: GGUF (compatible with llama.cpp, Ollama, LM Studio, GPT4All)
Base precision: BF16 intermediate, quantized to Q5_K_M

Training Details

Base Model

Model: Qwen/Qwen2.5-32B-Instruct
Architecture: Qwen2 (32B parameters)

Fine-Tuning

Method: QLoRA via Unsloth
LoRA rank: 32
LoRA alpha: 64
Target modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj
Max sequence length: 8,192 tokens
Precision: BF16 with 4-bit quantized base (QLoRA)
Optimizer: AdamW 8-bit
Learning rate: 2e-4 with cosine schedule
Epochs: 3 per adapter
Hardware: NVIDIA A100 80GB

Training Pipeline

Three-stage LoRA training, then merge and export:

Base Identity LoRA — 4,997 conversations covering all 16 executive personas, identity grounding, peer awareness, domain knowledge, autonomous behavior, and governance scenarios.
Personality LoRA — Long-form (25+ turn) conversations and drift-recovery data to maintain consistent persona over extended interactions.
Domain LoRAs (x4) — Specialist adapters trained on 50 conversations per executive per cluster, focused on deep domain reasoning.
Merge — Base + personality LoRAs merged into a single set of weights.
GGUF Export — Merged model and each domain adapter exported separately to Q5_K_M GGUF.

Training Data

Identity data: 1,259 training / 133 evaluation conversations (ShareGPT format)
Personality data: Long-form conversations (25+ turns) and drift-recovery scenarios
Domain data: 50 conversations per executive per cluster (800 total across 4 domains)
Sources: Claude (primary teacher), multi-model distillation (Qwen, Llama, Hermes, Mistral, Gemma, Phi, Command-R), synthetic generation
Supplemental: Adversarial governance, escalation scenarios, handoff routing, source literacy, production gap coverage

The 16 Executives

Role	Codename	Domain
Chief of Staff (CoS)	Overwatch	Coordination, delegation, orchestration
Chief Technology Officer (CTO)	Forge	Development, architecture, tech strategy
Chief Financial Officer (CFO)	Keystone	Finance, budgeting, fiscal analysis
Chief Marketing Officer (CMO)	Aegis	Marketing, brand, market positioning
Chief Information Officer (CIO)	Sentinel	IT governance, information security, compliance
Chief Product Officer (CPO)	Blueprint	Product strategy, roadmap, user experience
Chief Research Officer (CRO)	Axiom	Research methodology, data analysis, insights
Chief Data Officer (CDO)	Index	Data governance, analytics, data strategy
Chief Engineering Officer (CEngO)	Foundry	Engineering execution, DevOps, quality
Chief Security Officer (CSecO)	Citadel	Security operations, threat management, AppSec
Chief Customer Officer (CCO)	Echo	Customer experience, support, retention
Chief Strategy Officer (CSO)	Beacon	Corporate strategy, competitive intelligence
Chief Revenue Officer (CRevO)	Compass	Revenue operations, sales strategy, growth
Chief Risk Officer (CRiO)	Vector	Risk assessment, mitigation, regulatory
Chief Compliance Officer (CComO)	Accord	Compliance, policy, regulatory frameworks
Chief Operating Officer (COO)	Nexus	Operations, process optimization

Usage with Ollama

# Create the model
ollama create csuite-model -f Modelfile

# Run
ollama run csuite-model "What is your codename and domain?"

Each model includes a Modelfile with Qwen chat template (<|im_start|> / <|im_end|>), tuned inference parameters, and a default system prompt. The C-Suite runtime overrides the system prompt per-executive at inference time.

Inference Parameters (from Modelfiles)

Parameter	Base Model	Technical
temperature	0.7	0.6
top_p	0.9	0.9
top_k	40	40
num_ctx	8192	8192
repeat_penalty	1.1	1.1

Intended Use

These models are designed to be used as part of the C-Suite AI executive team platform. They are optimized for:

Executive persona role-playing with consistent identity
Domain-specific reasoning and analysis
Multi-agent collaboration and delegation
Autonomous decision-making within defined boundaries
Governance-aware behavior with appropriate escalation

Limitations

Models are fine-tuned for the C-Suite executive framework and may not generalize well to unrelated tasks
Domain specialists are trained for their specific cluster and may produce lower quality responses outside their domain
The merged base model covers all 16 executives but with less domain depth than the specialists
Models may occasionally reference internal system concepts (modules, tools) that only exist within the C-Suite runtime

License

This work is licensed under CC-BY-NC-4.0. You may share and adapt for non-commercial purposes with attribution. For commercial licensing, contact licensing@1450enterprises.com.

The base model (Qwen/Qwen2.5-32B-Instruct) is licensed under Apache 2.0 by the Qwen team at Alibaba Cloud.

Citation

@misc{csuite-executive-ai-2026,
  title={C-Suite Executive AI: Fine-Tuned Multi-Persona Language Models},
  author={Chris Arsenault},
  year={2026},
  publisher={1450 Enterprises LLC},
  url={https://github.com/chrisarseno/csuite-model}
}