Looking for an Open Source LLM Fine-Tuning & Evaluation Engineer — Paid Project

Xpert Systems is looking for an experienced Open Source LLM engineer/researcher for a focused paid project.

We have built hundreds of synthetic training and evaluation datasets, agentic datasets, digital twins, and synthetic worlds. We now want to independently demonstrate something very specific:

Does training on Xpert synthetic data measurably improve the capability of an open-weight LLM?

The project will be conducted entirely on AWS. We want to select an appropriate 7B–8B open-weight model, establish its performance against a frozen Xpert evaluation set, fine-tune it using an Xpert training dataset with LoRA/QLoRA, and rerun the exact evaluation.

The deliverable we ultimately want is easy and defensible:

Base Model → Baseline Score → Xpert Training Data → Fine-Tuned Model → Same Evaluation → Measured Improvement

We are looking for someone hands-on with AWS GPU/EC2, PyTorch, Hugging Face,



LoRA/QLoRA, open-weight LLMs, fine-tuning, and rigorous model evaluation. Experience building reproducible AI benchmarks and avoiding training/test contamination is particularly valuable.

This is not about producing an impressive demo. If the model doesn't improve, we want the experiment to tell us that too. We want scientifically credible results that we can reproduce across additional datasets and models.

If you've done this type of work and would like to collaborate on a paid pilot, please DM me or write to [email protected] with a short description of your LLM fine-tuning/evaluation experience and, ideally, links to relevant work.

#LLM #LLMTraining #FineTuning #MachineLearning #AIEngineering #SyntheticData #ModelEvaluation #LoRA #QLoRA #AWS #OpenSourceAI #AgenticAI

📌 Open Source LLM Fine-Tuning & Evaluation Engineer — Paid Project (India)
🏢 Xpert Systems
📍 India