AI Audio Specialist & Annotator | Project Hedgehog & Horizon
In this role, I specialized in training and evaluating advanced language models to improve their performance in audio engineering, music generation, and workflow comprehension. I provided structured feedback through Reinforcement Learning from Human Feedback (RLHF) by critiquing AI-generated music and documenting error traces. I enhanced model transparency by writing detailed metacognitive reasoning to explain technical corrections to audio outputs. • Evaluated generative audio samples for spectral integrity and phase coherence. • Authored 'Chain-of-Thought' documentation for audio corrections and workflow logic. • Improved LLM hardening by providing reproducible feedback on EQ, compression, and spatial audio tasks. • Focused on increasing dataset quality for reinforcement learning targeting music, sound design, and DAW-based workflows.