Teams at Mirage

Design and build end-to-end agentic systems for video-focused creative tasks. Develop training objectives, datasets, and fine-tuning strategies for LLMs, explore multimodal reasoning, run experiments and evaluations, analyze failure modes, and iterate to improve agent reliability and control in video analysis and editing workflows.

Train and optimize large-scale video and multimodal models; improve training and inference efficiency; apply distillation, quantization, and pruning; build distributed training systems; optimize GPU utilization and throughput; develop tooling for experimentation and debugging; productionize research models and monitor real-world performance.

Own full-cycle technical recruiting for Engineering and Product teams, source and engage passive talent, partner with hiring managers, negotiate offers, and improve hiring processes and candidate experience while building diverse pipelines.