Remote LLM Fine-Tuning Engineer for Production

Workana; Mexico; Hybrid; PHP 3,743,000 - 7,486,000; Part time

Job summary

medxprts.ai is seeking an AI Engineer - LLM Fine-Tuning to personally train and fine-tune open-weight models, build training infra, and ship results into production. This fully remote position starts part-time with potential to transition to full-time and requires overlap with US timezones.

You will design evaluation harnesses, handle domain-specific data, and collaborate with engineers to optimize inference and deployment while maintaining HIPAA-awareness when handling sensitive data.

Qualifications

• Experience fine-tuning open-weight LLMs with LoRA/QLoRA or full SFT and shipped examples.
• Experience building training infrastructure using PyTorch and Hugging Face tooling.
• Dataset engineering for unstructured documents; deduplication and contamination checks.

Responsibilities

• Design and run LLM fine-tuning experiments on open-weight models and ship trained models into product workflows.
• Build training infrastructure with multi-GPU training orchestration on AWS or GCP.
• Engineer datasets from long PDFs and medical/legal records, including deduplication and train/eval splits.
• Create evaluation harnesses with held-out tests and human-calibrated judging.
• Implement preference/feedback training workflows (DPO/RLHF) and integrate into retraining pipelines.
• Collaborate with backend/frontend engineers to integrate fine-tuned models into services and optimize latency/cost.
• Participate in PR reviews, releases, and ensure secure handling of sensitive data.

Skills

LLM fine-tuning

PyTorch

Hugging Face tooling

Dataset engineering

Evaluation and calibration

RLHF/DPO workflows

Production deployment

English communication

Tools

Transformers

PEFT

TRL/Axolotl

DeepSpeed

AWS

GCP

Git

Job description

medxprts.ai is seeking an AI Engineer - LLM Fine-Tuning to personally train and fine-tune open-weight models, build training infra, and ship results into production. This fully remote position starts part-time with potential to transition to full-time and requires overlap with US timezones.

You will design evaluation harnesses, handle domain-specific data, and collaborate with engineers to optimize inference and deployment while maintaining HIPAA-awareness when handling sensitive data.