Research Scientist Model Efficiency

Skills

About the Role

Develop methods that make AI models cheaper and faster to serve while preserving important quality, including quantization, sparsity, pruning, speculative decoding, MTP, attention, and KV-cache techniques.

Requirements

• Bachelor’s, Master’s, or PhD in Computer Science, Electrical Engineering, or a related field
• Hands-on experience in LLM inference, model optimization, or ML systems
• Strong Python programming ability and deep familiarity with PyTorch
• Implementation-level depth in at least one model-efficiency area
• Strong understanding of transformer internals
• Rigorous evaluation experience with task-level metrics, controlled comparisons, and honest baselines
• Experience taking efficiency methods into production serving or equivalent research depth preferred
• Familiarity with vLLM, SGLang, or TensorRT-LLM preferred
• Publications at top-tier systems or ML venues or substantial open-source contributions welcome
• Enthusiasm for AI infrastructure and efficient inference with strong ownership and engineering discipline

Responsibilities

• Implement and adapt published model-efficiency methods on models and hardware
• Develop new model optimizations
• Evaluate throughput and quality claims using rigorous task-level metrics and controlled comparisons
• Analyze accuracy loss in model behavior
• Apply efficiency methods to production serving

Benefits

• Welfare benefits
• Training and mentoring
• Inclusive work environment
• Fast-growing company with networking opportunities
• Personal accountability, autonomy, growth, and learning opportunities