LLM Data Science Manager — AI Model Training Lead

Lexisnexis Legal & Professional®; California; On-site; USD 115,000 - 192,000; Full time

Benefits offered by this job

Annual incentive bonus

Country-specific benefits

Job summary

LexisNexis Legal & Professional® is seeking a Manager Data Science to lead a team of data scientists in the design, training, and fine-tuning of large language models for business-critical tasks. The role emphasizes best-practice governance, reproducible pipelines, and collaboration with product and engineering to deploy models.

The candidate will oversee data preparation, experimentation, and evaluation, while driving efficiency across multi-GPU environments and supporting model deployment

Qualifications

• Strong understanding of transformer architectures and language modeling objectives.
• Experience implementing supervised fine-tuning (SFT) and tuning hyperparameters.
• Knowledge of parameter-efficient fine-tuning (PEFT) techniques like LoRA/QLoRA.
• Experience with multi-GPU training and distributed strategies such as FSDP or DeepSpeed.

Responsibilities

• Lead the design and execution of LLM training and fine-tuning projects.
• Oversee preparation of training datasets including cleaning and annotation.
• Develop and optimize fine-tuning workflows and evaluation methods.
• Establish benchmarks for factual accuracy, safety, and domain relevance.
• Diagnose training issues and optimize GPU utilization and efficiency.
• Manage and mentor data scientists and ensure reproducible practices.
• Collaborate with product and engineering teams for deployment needs.

Skills

LLM fundamentals

Python

PyTorch

Hugging Face

SFT training

PEFT/Lora

Tools

Hugging Face Transformers

Datasets

PyTorch FSDP

DeepSpeed

Job description

LexisNexis Legal & Professional® is seeking a Manager Data Science to lead a team of data scientists in the design, training, and fine-tuning of large language models for business-critical tasks. The role emphasizes best-practice governance, reproducible pipelines, and collaboration with product and engineering to deploy models.

The candidate will oversee data preparation, experimentation, and evaluation, while driving efficiency across multi-GPU environments and supporting model deployment