Work remotely or hybrid in the Greater Toronto Area, adapting open-weight large language models with PyTorch and Hugging Face. Build measured results from datasets, training, and evaluation.

As a Fall/Winter Software Intern, you will work hands-on with real models and GPUs alongside the founding team. You will prepare datasets, fine-tune models, build benchmarks, and compare training runs. The work applies AI to security and identity.

Key Responsibilities:

- Collect, clean, and prepare training datasets
- Prepare evaluation datasets for model assessment
- Fine-tune open-weight LLMs using SFT, LoRA, and QLoRA
- Run and monitor GPU training jobs
- Compare runs and report what worked

Requirements:

- PyTorch and Hugging Face Transformers experience
- Experience with Datasets and PEFT
- Hands-on open-weight LLM experience
- Training and evaluation dataset preparation experience
- Understanding of GPU memory, batch sizes, and mixed precision

Apply PyTorch, Hugging Face, and GPU fundamentals to evaluate models for security and identity work.

#J-18808-Ljbffr

📌 Large Language Model Software Intern (Toronto)
🏢 Arkion Identity Systems
📍 Toronto