ML Engineer at Turing USA
Prompt + Response Writing (SFT)Performed supervised fine-tuning (SFT) training work for the OS-World project, preparing the system to execute tasks in a Linux environment. Worked on coding/problem-solving datasets for C++ tasks to identify issues and apply solution optimizations. Contributed to iterative improvements of a Google project’s Git-based responses by comparing performance against competitors. • Conducted SFT training for Linux task performance (OS-World) • Analyzed C++ Codeforces solutions, found issues, and optimized model outputs • Improved Git-based response quality in the Google Antigravity project • Worked on Bash-based problems for Terminal Bench with goals to enhance commercial model performance
2025 - Present