Teams at G2i

Evaluate AI coding agents such as Codex, Claude Code, and Cursor by assessing reasoning, explanations, engineering judgment, and overall interaction quality. Provide detailed written and recorded feedback, distinguish response quality levels, and help define standards for excellent AI-assisted development workflows. This flexible, remote contract requires strong software engineering experience, familiarity with modern AI development tools, rigorous subjective judgment, and excellent English communication.

Lead the company’s Physical AI research and engineering function as a hands-on technical leader. Define research priorities, design and train Transformer-based models, establish rigorous offline and real-world evaluations, integrate systems with physical robots, and analyze failures across data, models, hardware, and environments. Build a focused team, guide technical architecture and experiments, translate customer needs into requirements, and deliver measurable improvements through deployed models, benchmarks, evaluation systems, or data products.