Intern for a team that builds and fine-tunes LLMs for banking scenarios. Tasks include training models on GPU clusters, preparing datasets, analyzing experiments, and optimizing inference.
Tasks:
- Train LLMs on GPU clusters with H100
- Collect, clean, and prepare datasets for pre-training and fine-tuning
- Analyze experiment results and improve model quality
- Run existing and create new benchmarks
- Optimize inference, conduct load testing, and deploy trained models
Requirements:
- Final year undergraduate, Master's, or PhD student at a technical university
- Full-time studies
- Deep understanding of theoretical and practical aspects of LLMs: architecture, tokenization, pipeline construction, batching, mixed precision, distributed training
- Proven practical experience in training LLMs (pre-training, fine-tuning, alignment) in academic projects, papers, or courses
- Python development experience, Linux experience, knowledge of standard bash commands
- Attention to detail
Conditions:
- Paid internship
- Internship duration: 3 months, 40 hours per week from 9:00 to 18:00
- Hybrid work format with at least 3 days per week in the office
- Comfortable modern office located at: Moscow, Kutuzovsky Prospekt 32
Skills:
- Python
- Linux
- bash
- LLM
- GPU
- H100
- pre-training
- fine-tuning
- alignment
- batching
- mixed precision
- Weights & Biases
- MLflow
- Langfuse
