Responsibilities

• Build datasets through scrappy methods, including:
  • Synthetic data pipelines
  • Data scrapers
  • Combining open source datasets
  • Spinning up data annotations
• Conduct experiments to find good data mixes, regularisers, and hyperparameters
• Implement and test different fine-tuning and preference learning techniques
• Own experiments end-to-end, including:
  • Data collection and cleaning
  • Implementing scalable data pipelines
  • Designing processes and software to facilitate ML experimentation
  • Implementing and debugging new ML frameworks and approaches
  • Training models
  • Building tooling to evaluate and play with models
• Improve various aspects such as:
  • Data quality
  • Data formatting
  • Job startup speed
  • Evaluation speed
  • Ease of experimentation
• Adjust infrastructure for model inference, including improving constrained generation for tool-use
• Work with engineering to integrate models into Proxy

Requirements

• Direct experience training LLMs or VLMs with methods such as:
  • Distillation
  • Supervised fine-tuning
  • Policy optimisation
• Experience with large-scale distributed training and inference
• Experience debugging ML systems and codebases
• Proficiency in frameworks like PyTorch
• Strong general foundations in software engineering, with an interest in non-ML software

Bonus Qualifications

• Experience with open source models
• Experience with frameworks for fine-tuning and RLHF
• Familiarity with public datasets (including synthetic ones)

Why Join Us?

• Be at the cutting edge of AI and foundation models
• Work on challenging problems that impact users' daily lives
• Collaborative and innovative work environment
• Opportunities for professional growth and learning
• Competitive salary plus equity
• Benefits include:
  • 30 days PTO
  • Private Medical Cover
  • Pension
  • Wellness Benefit
  • Lunch Allowance