Language Model Training Intern

Join Razer's global team to revolutionize gaming.

Job Responsibilities: Train and optimize language models for Razer Software across the full pipeline , pretraining corpus construction, fine-tuning, evaluation, and deployment optimization.

Job Scope Work with senior data scientists and engineers on the model powering Razer Software, covering:

• Corpus construction , cleaning, filtering, deduplication, and mixing of training data
• Fine-tuning , SFT, LoRA, and preference optimization runs
• Evaluation , quality, latency, and memory benchmarking against production constraints
• Deployment optimization , compression and quantization experiments for on-device targets

Learning Objectives By the end of this internship, you will have:

• Hands-on experience across the full LLM training lifecycle at production scale
• Practical judgment in data engineering: how cleaning/filtering/mixing decisions propagate into model quality
• The ability to independently run a rigorous ML experiment loop , hypothesize, train, evaluate, diagnose subtle regressions, iterate
• Direct exposure to how deployment constraints shape training and architecture decisions
• Mentorship from senior engineers and visibility into how a production roadmap for a shipping AI feature actually gets decided

Candidate Requirements

• Education: Current Bachelor's, Master's, or PhD student in CS, AI, Data Science, or a related field
• Must-have knowledge: LLM training paradigms, Transformer/deep learning fundamentals, how deployment constraints shape training decisions
• Must-have skills: Python, hands-on PyTorch model training, data engineering for training corpora, end-to-end experiment running, benchmarking quality/latency/memory
• Nice-to-have: Model compression, distributed training, Hugging Face/Accelerate/DeepSpeed, Linux environment
• Screening bar: Demonstrable hands-on model training/fine-tuning experience
• Strong positives: Trained a model from scratch, multi-GPU training, model compression/on-device deployment work, top-tier publications