LLM Engineer - AI-Assisted RTL Integration

Job Overview:
LLM Engineer with expertise in prompt engineering, dataset creation, fine-tuning, and deployment of on-premises open-source LLMs for RTL (Register Transfer Level) design. The ideal candidate will work closely with RTL domain experts to develop and optimize AI-assisted RTL integration workflows.
The role involves prompt engineering, output validation and re-prompting, fine-tuning of LLMs when necessary, and building datasets to enhance model accuracy using the latest AI/ML technologies.
Key Responsibilities:
1. LLM Deployment & Integration:

• Deploy and optimize open-source LLMs on-premises for RTL integration.
• Develop custom pipelines for LLM-assisted RTL design, analysis, and verification.
• Work with RTL experts to fine-tune prompts for best performance.

2. Prompt Engineering & Optimization:

• Design, refine, and test effective prompts for RTL integration tasks.
• Continuously evaluate LLM responses and develop strategies for output validation.
• Implement re-prompting techniques to improve accuracy and efficiency.

3. Dataset Creation & Fine-Tuning:

• Identify gaps in model accuracy and develop datasets for model retraining.
• Collect, clean, and curate RTL-specific datasets to improve model performance.
• Fine-tune LLMs using state-of-the-art training frameworks (e.g., PyTorch, TensorFlow, Hugging Face).
• Experiment with latest AI/ML techniques to optimize LLM efficiency for RTL workflows.

4. Model Validation & Performance Tuning:

• Implement evaluation metrics to measure model performance in RTL tasks.
• Conduct benchmarking and performance tuning to ensure model accuracy.
• Develop feedback loops for continuous improvement of LLM-assisted RTL processes.

5. Collaboration & Research:

• Work closely with RTL engineers to understand domain challenges.
• Stay up to date with the latest advancements in AI/ML and hardware design automation.
• Evaluate and implement state-of-the-art LLM architectures for RTL-specific applications.

Key Skills

• Strong expertise in LLMs - Open-source models like Llama, Falcon, Mistral, or GPT-based architectures.
• Experience in fine-tuning LLMs using PyTorch, TensorFlow, or Hugging Face.
• Prompt engineering expertise - Ability to craft optimized prompts for RTL tasks.
• Understanding of RTL design (basic knowledge preferred) and how AI can assist in hardware workflows.
• Experience in data collection, preprocessing, and synthetic dataset creation for fine-tuning.
• Familiarity with LLM inference optimization techniques (e.g., quantization, pruning).
• Strong coding skills in Python and frameworks like LangChain, LlamaIndex, or OpenAI APIs.
• Experience with containerization (Docker, Kubernetes) and on-prem model hosting.