🛠 Tech Stack

💼 About This Role

You'll build and improve core ML systems for training specialized AI models at a well-funded startup. Your work will directly shape custom model quality at scale, measured by frontier performance and efficiency. You'll own the full training lifecycle from data intake to shipped results.

🎯 What You'll Do

• Lead projects from data intake through the full training pipeline
• Build and maintain data processing pipelines for training data
• Create dashboards and visualization tools for training metrics
• Train models using internal frameworks and iterate on evaluations

📋 Requirements

• 2+ years experience training AI models using PyTorch
• Hands-on experience with post-training LLMs using SFT or RL
• Strong understanding of transformer architectures and training
• Experience with LLM-specific training frameworks (e.g., Hugging Face Transformers, DeepSpeed, Axolotl)

✨ Nice to Have

• Experience with model distillation or knowledge transfer
• Experience building dashboards and data visualization tools
• Familiarity with vision encoders and multimodal models

🎁 Benefits & Perks

• 💰 Competitive compensation and equity
• 🏢 In-person office in downtown San Francisco
• 🚀 High-growth startup environment
• 🖥️ Large compute budget and GPU reservation

📨 Hiring Process

Send resume and GitHub to [email protected] or apply via Ashby.

[email protected]