Research Engineer, Agentic Systems

About the role

Mirage is looking for a Research Engineer to advance large language models for multimodal creative tasks, especially video. You will design and build agentic systems that can understand and act on complex, real‑world data, pushing the boundaries of AI‑driven video production.

Key responsibilities

• Design and implement end‑to‑end agentic pipelines for creative video tasks.
• Develop novel training, fine‑tuning, and alignment methods for large language models.
• Create new objectives, datasets, and evaluation metrics to improve agent reliability.
• Explore multimodal reasoning and structured generation for creative control.
• Run systematic experiments and build evaluation frameworks for video analysis and editing.
• Analyze failure modes across planning, tool use, and execution, iterating on improvements.

Required profile

• BS/MS/PhD in Computer Science, Machine Learning, or a related field.
• Proven experience building production‑grade ML systems or agentic pipelines.
• Deep understanding of transformers and modern LLM techniques.
• Hands‑on experience with fine‑tuning, alignment, or post‑training methods for structured output generation.
• Comfort owning the full stack from model experiments to deployed agents.
• Strong experimental rigor and practical sense of what makes agents work.

Required skills

• Transformers
• Large Language Models (LLMs)
• Fine‑tuning
• Model alignment
• Post‑training methods
• Structured output generation
• Tool‑use integration
• Production ML systems
• Agentic pipelines
• Multimodal reasoning
• Video analysis

What we offer

• Comprehensive medical, dental, and vision plans
• 401(k) with employer match
• Commuter benefits
• Catered lunches