Senior Remote LLM Fine-Tuning Engineer

Bright Vision Technologies; United States; Remote; USD 100,000 - 150,000; Full time

Job summary

Bright Vision Technologies is seeking a Large Language Model Specialist for 100% remote work in the United States. This full-time direct W2 role requires 6+ years of experience and pays $100,000 to $150,000 annually.

You will design, implement and operate large-scale fine-tuning workflows across supervised, DPO, and RLHF approaches. Responsibilities include dataset construction, scalable training pipelines, and rigorous evaluation.

Qualifications

• Master’s or PhD in CS/ML or equivalent experience.
• 6+ years of ML research and engineering with LLM exposure.
• Strong Python and PyTorch proficiency.
• Hands-on fine-tuning of transformer models at scale.
• Familiar with distributed training (FSDP/ZeRO) and RLHF/DPO.
• Strong evaluation design and methodology.

Responsibilities

• Design and run fine-tuning experiments for LLMs (supervised, DPO, RLHF).
• Lead dataset construction, curation and QA for instruction data.
• Build scalable training pipelines on modern frameworks.
• Tune hyperparameters and stability strategies for large models.
• Implement LoRA/QLoRA and adapters.
• Design rigorous evaluation suites and benchmarks.
• Collaborate with product, research and platform teams.
• Mentor junior engineers and share best practices.

Skills

Python

PyTorch

LLM Fine-Tuning

Distributed Training

RLHF/DPO

Evaluation Design

Communication

Production Engineering

Education

Master’s/PhD in CS/ML or equivalent

Tools

FSDP

ZeRO

Pipeline Parallelism

Job description

Bright Vision Technologies is seeking a Large Language Model Specialist for 100% remote work in the United States. This full-time direct W2 role requires 6+ years of experience and pays $100,000 to $150,000 annually.

You will design, implement and operate large-scale fine-tuning workflows across supervised, DPO, and RLHF approaches. Responsibilities include dataset construction, scalable training pipelines, and rigorous evaluation.