AI Engineer, Algorithm Evaluation & Agentic Systems

What you'll do

• Design and scale comprehensive evaluation pipelines for image and video understanding tasks.
• Conduct deep failure analysis to identify root causes of hallucinations and temporal inconsistencies in video models.
• Build and deploy agentic workflows that use vision models to solve multi-step user problems.
• Curate high-quality, schematized datasets and ground-truth benchmarks for multi-modal capabilities.
• Provide data-driven insights and metrics to model training teams to guide fine-tuning efforts.
• Develop robust metrics and automated regression testing for generative AI and foundation models.
• Implement LLM/VLM-as-a-judge methodologies and human-in-the-loop evaluation systems.

What we're looking for

• Master's degree and at least 3 years of relevant industry experience.
• At least 3 years of applied experience in Machine Learning, Computer Vision, or AI System Evaluation.
• Deep understanding of core ML principles including probability, statistics, data distributions, and model bias/variance.
• Theoretical and practical expertise in Computer Vision (CV) and Vision-Language Models (VLMs).
• Proven track record of designing robust evaluation frameworks for generative AI or foundation models.
• Experience building and evaluating LLM/VLM-powered agents including tool use, reasoning, and memory management.
• Proficiency in Python and experience with PyTorch for inference and building scalable pipelines.
• Ability to perform deep failure analysis to identify hallucinations and performance bottlenecks.