AI Engineer – LLaMA 3 Agent Integration (ERP + SQL)

AI Engineer

Key Responsibilities:

• Setup and deploy

  LLaMA 3 (8B/FP16) models

  using llama-cpp-python or Hugging Face
• Integrate the AI model with

  FastAPI

  to create secure REST endpoints
• Connect with

  MS SQL database

  and design query logic for ERP modules (Sales, Payments, Units, etc.)
• Implement

  prompt engineering

  or

  fine-tuning (LoRA)

  to improve SQL generation accuracy
• Build a user-facing interface (React or basic web UI) for interacting via

  text or voice

• Integrate

  Whisper (OpenAI)

  or any STT system to support voice commands
• Ensure model responses are secure, efficient, and auditable (only SELECT queries allowed)
• Supervise or perform

  supervised fine-tuning

  with custom ERP datasets
• Optimize for

  performance (GPU usage)

  and

  accuracy (prompt/RAG tuning)

Must-Have Skills:

• Strong experience with

  LLM deployment

  (LLaMA 3, Mistral, GPT-type models)
• Solid Python development experience using

  FastAPI

  or Flask
• SQL knowledge (esp.

  MS SQL Server

  ) – must know how to write and validate queries
• Experience with

  llama-cpp-python

  , Hugging Face Transformers, and LoRA fine-tuning
• Familiarity with

  LangChain

  or similar LLM frameworks
• Understanding of

  Whisper (STT)

  or equivalent Speech-to-Text tools
• Experience working with

  GPU inference (NVIDIA 4070/5090 etc.)