Description

You will leverage large language models and advanced NLP techniques to extract insights from large textual datasets.

Responsibilities

• Integrate offline LLMs such as LLaMA2/3 and Mistral to meet project-specific goals.
• Fine-tune pre-trained models using PEFT techniques like LoRA and QLoRA to align with enterprise requirements.
• Engineer features to capture linguistic nuances and optimize model performance.
• Implement RAG techniques and optimize models for efficiency, coherence, and relevance.
• Evaluate LLM performance and implement optimizations for accuracy and text generation tasks.

Required Skills

• 16 to 20 years of experience in data science and machine learning.
• Proficiency in Python for data analysis, statistical modeling, and machine learning.
• Hands-on experience with Huggingface, TensorFlow, and PyTorch.
• Experience with fine-tuning infrastructure, including managing GPU memory and compute for various model sizes.
• Deep understanding of SLMs and Tiny LLMs like phi3 and OpenELM.
• Knowledge of fine-tuning recipes including next token prediction and fill-in-the-middle.
• Expertise in data manipulation using Pandas and NumPy.
• Strong understanding of statistical techniques and data analysis.
• Ability to communicate technical concepts to non-technical audiences.

Key Skills

Education

Any Gradute