Outlier
I have contributed to advanced data training and AI evaluation projects on Outlier, specifically working across specialized pipelines like Antechamber, Icarus, Phoenix, and Iris Gambit. Rather than handling basic microtasks, my role involved crafting complex engineering prompts, performing deep side by side response evaluations, and catching edge case bugs in code. Through these coding and advanced technical tracks, I focused on optimizing model logic, evaluating algorithmic reasoning, and fixing complex flaws before these AI models are shipped to millions of end users.