AI Pipeline Engineer: LoRA Fine-Tuning in Production

Optime; Austin (TX); Hybrid; USD 140,000 - 210,000; Full time

Benefits offered by this job

Flexible remote work

Competitive compensation

Equity potential

Job summary

Optime is building next-generation AI infrastructure and seeks an AI/ML Engineer (Python) to productize fine-tuning techniques (LoRA/QLoRA) for vendors.

You will design training pipelines using PyTorch and the Hugging Face ecosystem, automate data preprocessing, and integrate with Node.js backends and job queues.

This role emphasizes reproducible training environments, testing adapter outputs, and occasional in-office collaboration.

Qualifications

• Strong Python programming skills.
• Hands-on PyTorch and Hugging Face ecosystem experience.
• Familiarity with LoRA/QLoRA or PEFT methods.

Responsibilities

• Implement and maintain LoRA/QLoRA fine-tuning pipelines.
• Develop incremental training and adapter stacking workflows.
• Automate data preprocessing for user datasets.
• Build training scripts and workflows with Node.js backends and job queues.
• Add monitoring hooks for dashboards.
• Collaborate with DevOps for reproducible environments.
• Write tests ensuring reproducibility of adapter outputs.
• Willingness to occasionally be present in the office for discussions.

Skills

Python

Linux

Production-grade software

Team collaboration

Problem solving

Tools

PyTorch

Hugging Face Transformers

PEFT

Datasets

CUDA

REST/gRPC

Node.js integration

NCCL

DeepSpeed

Accelerate

bitsandbytes

xformers

flash-attention

Job description

Optime is building next-generation AI infrastructure and seeks an AI/ML Engineer (Python) to productize fine-tuning techniques (LoRA/QLoRA) for vendors.

You will design training pipelines using PyTorch and the Hugging Face ecosystem, automate data preprocessing, and integrate with Node.js backends and job queues.

This role emphasizes reproducible training environments, testing adapter outputs, and occasional in-office collaboration.