adamabuhamdan/startup-advisor-dataset

🚀 Startup Advisor Dataset

A high-quality instruction-following dataset distilled from 8 foundational business and startup books, structured as actionable advice with real-world 2025 examples. Designed for fine-tuning large language models (e.g., Qwen, LLaMA, Mistral) to become expert startup advisors.

---

📖 Dataset Summary

Property	Value
Total Entries	1,564
Format	JSONL — ChatML (messages array)
Language	English
License	CreativeML OpenRAIL-M
Avg. Tokens / Entry	~288
Total Tokens (est.)	~451,000

---

🗂️ Categories & Distribution

Category	# Entries	Source Books
📈 Scaling	400	Blitzscaling, Measure What Matters
💰 Funding	347	The Angel Investor's Handbook, Venture Deals
👥 Team & Leadership	253	Leaders Eat Last, The Hard Thing About Hard Things, Rules
🌱 Growth & Traction	146	Crossing the Chasm, Traction
💡 Ideation	135	Zero to One, Blue Ocean Strategy, The Mom Test
🏆 Strategic Maturity	111	Blue Ocean Strategy, Good Strategy Bad Strategy
✅ Validation	86	The Lean Startup, Sprint
🛠️ Product Building	86	Hooked, Lean Product Playbook

---

📐 Data Format

Each entry follows the ChatML instruction-following format with 3 roles: system, user, and assistant. A _meta field provides structured metadata for filtering (ignored during training).

{
  "messages": [
    {
      "role": "system",
      "content": "You are an expert startup and business advisor. Provide concise, actionable advice grounded in proven frameworks and real-world examples from leading companies."
    },
    {
      "role": "user",
      "content": "I'm at the **Scaling** stage of building my startup.\nGive me actionable advice on **blitzscaling speed**.\nRelevant themes: speed, growth, market_capture.\nComplexity level: advanced."
    },
    {
      "role": "assistant",
      "content": "Prioritize speed over efficiency when your market window is closing. Accept chaos, technical debt, and management strain as temporary costs of capturing a winner-take-all position before rivals do.\n\n**Why this matters:** In fast-moving markets, the first company to scale dominates distribution, talent, and network effects — making catch-up nearly impossible.\n\n**Real-world example (2025):** OpenAI blitzscaled its API platform in 2025, accepting early instability to lock in developers before Google and Anthropic consolidated the market.\n\n**Avoid:** `optimizing_too_early`, `waiting_for_perfect_infrastructure`, `prioritizing_margins_over_growth`"
    }
  ],
  "_meta": {
    "advice_id": "advice_scaling_001",
    "source_file": "Blitzscaling.md",
    "category": "Scaling",
    "stage": "Scaling",
    "topic": "blitzscaling_speed",
    "complexity": "advanced",
    "tags": ["speed", "growth", "market_capture"]
  }
}

---

🧠 Complexity Distribution

Each entry is labeled with one of three complexity levels to support curriculum learning:

Level	Description
beginner	Core concepts, widely applicable principles
intermediate	Nuanced trade-offs, context-dependent decisions
advanced	Deep strategy, founder-level judgment calls

---

🔧 Usage

Load with Hugging Face datasets

from datasets import load_dataset

ds = load_dataset("adamabuhamdan/startup-advisor-dataset", split="train")
print(ds[0])

Fine-tune with trl SFTTrainer (Qwen / LLaMA)

from trl import SFTTrainer, SFTConfig
from datasets import load_dataset
from transformers import AutoTokenizer, AutoModelForCausalLM

model_id = "Qwen"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model     = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype="auto")

dataset = load_dataset("adamabuhamdan/startup-advisor-dataset", split="train")

trainer = SFTTrainer(
    model=model,
    tokenizer=tokenizer,
    train_dataset=dataset,
    args=SFTConfig(
        output_dir="./startup-advisor-ft",
        num_train_epochs=3,
        per_device_train_batch_size=2,
        gradient_accumulation_steps=8,
        learning_rate=2e-4,
        bf16=True,
    ),
)
trainer.train()

Filter by category or complexity

ds = load_dataset("adamabuhamdan/startup-advisor-dataset", split="train")

# Keep only advanced Scaling entries
import json
advanced_scaling = [
    row for row in ds
    if row["_meta"]["category"] == "Scaling"
    and row["_meta"]["complexity"] == "advanced"
]

---

📚 Source Books

Book	Author	Category
Zero to One	Peter Thiel	Ideation
The Mom Test	Rob Fitzpatrick	Ideation
Blue Ocean Strategy	Kim & Mauborgne	Ideation, Strategic Maturity
The Lean Startup	Eric Ries	Validation
Sprint	Jake Knapp	Validation
Hooked	Nir Eyal	Product Building
Lean Product Playbook	Dan Olsen	Product Building
Crossing the Chasm	Geoffrey Moore	Growth & Traction
Traction	Gabriel Weinberg	Growth & Traction
The Angel Investor's Handbook	Various	Funding
Venture Deals	Brad Feld	Funding
Blitzscaling	Reid Hoffman	Scaling
Measure What Matters	John Doerr	Scaling
Leaders Eat Last	Simon Sinek	Team & Leadership
The Hard Thing About Hard Things	Ben Horowitz	Team & Leadership
Good Strategy / Bad Strategy	Richard Rumelt	Strategic Maturity

---

⚠️ Limitations & Bias

• All real-world examples reference 2025 AI-era companies (OpenAI, Vercel, Notion, Figma, Perplexity, etc.) — the model may overfit to this era.
• Advice reflects frameworks from Western, VC-backed startup culture and may not generalize to all market contexts.
• The dataset does not contain negative examples or failure-mode demonstrations beyond avoid fields.

---

📄 Citation

@dataset{abuhamdan2025startup,
  author    = {Adam Abu Hamdan},
  title     = {Startup Advisor Dataset},
  year      = {2025},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/datasets/adamabuhamdan/startup-advisor-dataset}
}

---

Built with ❤️ to make world-class startup wisdom accessible through AI.