Skip to content

DEEPSEEK_SETUP.md

Latest commit

 

History

History
284 lines (204 loc) · 5.35 KB

DEEPSEEK_SETUP.md

File metadata and controls

284 lines (204 loc) · 5.35 KB
Raw

Use DeepSeek with Neuro

🚀 Why DeepSeek?

DeepSeek is an excellent choice for powering Neuro:

✅ Cost-Effective - Often cheaper than OpenAI
✅ High Performance - Competitive with GPT-4
✅ Fast - Low latency responses
✅ OpenAI-Compatible API - Easy to integrate

⚡ Quick Setup (3 Steps)

1. Get DeepSeek API Key

  1. Visit: https://platform.deepseek.com
  2. Sign up / Log in
  3. Go to API Keys
  4. Create new API key
  5. Copy the key (starts with sk-...)

2. Set Environment Variable

Windows (PowerShell):

$env:DEEPSEEK_API_KEY="sk-your-key-here"

Windows (Command Prompt):

set DEEPSEEK_API_KEY=sk-your-key-here

Mac/Linux:

export DEEPSEEK_API_KEY=sk-your-key-here

Permanent (add to .env file):

echo "DEEPSEEK_API_KEY=sk-your-key-here" > .env
pip install python-dotenv

3. Run Neuro

python neuro my_job_search.neuro

You'll see:

✓ AI parsing enabled with DeepSeek

That's it! AI is now active! 🎉

📊 DeepSeek vs OpenAI

Feature DeepSeek OpenAI GPT-4
Cost ~$0.14 per 1M tokens ~$30 per 1M tokens
Speed Fast Fast
Quality Excellent (comparable to GPT-4) Excellent
API OpenAI-compatible Native
Privacy China-based US-based

DeepSeek is ~200x cheaper! 💰

🎯 What AI Enables

Once DeepSeek is configured, you can use natural language:

Before (Without AI):

// Must follow exact syntax
pipeline FindJobs {
    goal: "Find AI engineer jobs"
    target_roles: ["ai engineer", "ml engineer"]
    locations: ["remote", "US"]
}

After (With DeepSeek):

// Natural language works!
goal: "yo find me some remote ai jobs, i know python and pytorch"

DeepSeek understands:

  • "yo" = casual greeting
  • "find me" = search action
  • "remote ai jobs" = remote AI engineer positions
  • "i know python and pytorch" = skills filter

Neuro automatically:

  • Searches job boards
  • Filters for Python/PyTorch requirements
  • Prioritizes remote positions
  • Generates report

💡 Example Usage

Input:

"I'm transitioning from QA to AI engineering, find me entry-level remote positions"

DeepSeek Parses:

{
  "intent_type": "job_search",
  "target_roles": ["ai engineer"],
  "experience_level": "entry-level",
  "locations": ["remote"],
  "inferred": {
    "current_role": "qa engineer",
    "career_transition": true,
    "needs": ["resume_help", "skill_gap_analysis"]
  }
}

Neuro Executes:

  • ✓ Searches for entry-level AI jobs
  • ✓ Prioritizes remote positions
  • ✓ Generates transition-focused resume suggestions
  • ✓ Provides skill recommendations

🔧 Technical Details

How DeepSeek Integration Works:

from openai import OpenAI

# DeepSeek uses OpenAI-compatible API
client = OpenAI(
    api_key=os.getenv('DEEPSEEK_API_KEY'),
    base_url="https://api.deepseek.com"
)

# Use deepseek-chat model
response = client.chat.completions.create(
    model="deepseek-chat",
    messages=[
        {"role": "system", "content": "Parse Neuro intent"},
        {"role": "user", "content": goal}
    ]
)

It's that simple! Same OpenAI client library, different endpoint.

💰 Cost Comparison

Typical Neuro Usage:

Per Job Search Run:

  • Parse intent: ~1,000 tokens
  • Cost with DeepSeek: ~$0.0001
  • Cost with OpenAI: ~$0.02

Monthly (30 runs):

  • DeepSeek: ~$0.003 (basically free!)
  • OpenAI: ~$0.60

DeepSeek is 200x cheaper! 🎯

🌟 Priority Order

Neuro checks for API keys in this order:

  1. DEEPSEEK_API_KEY - Preferred (cheap, fast)
  2. OPENAI_API_KEY - Fallback (more expensive)
  3. Pattern matching - No API needed (free, works offline)

Set DeepSeek first for best value!

🔒 Security

Store API Keys Safely

❌ Don't:

api_key = "sk-1234..."  # Hard-coded in code

✅ Do:

# Environment variable
export DEEPSEEK_API_KEY=sk-...

# Or .env file
DEEPSEEK_API_KEY=sk-...

For GitHub Actions: Add as repository secret:

  • Settings → Secrets → Actions → New secret
  • Name: DEEPSEEK_API_KEY
  • Value: Your API key

🎓 Getting Started

Option 1: DeepSeek (Recommended - Cheap!)

# 1. Get key from https://platform.deepseek.com
# 2. Set environment variable
export DEEPSEEK_API_KEY=sk-your-key

# 3. Install package
pip install openai

# 4. Run
python neuro my_task.neuro

Option 2: OpenAI (Alternative)

# 1. Get key from https://platform.openai.com
# 2. Set environment variable
export OPENAI_API_KEY=sk-your-key

# 3. Install package
pip install openai

# 4. Run
python neuro my_task.neuro

Option 3: No AI (Free)

# No setup needed!
# Works offline
# Use structured syntax

python neuro my_task.neuro

📚 More Information

🎯 Summary

DeepSeek Setup:

  1. Get API key: https://platform.deepseek.com
  2. Set: DEEPSEEK_API_KEY=sk-...
  3. Run: python neuro my_task.neuro

Cost: ~$0.0001 per run (basically free!)

Benefit: Natural language goals, smart inference, helpful clarifications

Perfect for Neuro! 🚀