AI Engineer | Scale AI (Remote)
Performed instruction tuning and reinforcement learning workflows to improve large language model behavior for engineering, data science, and math use cases. Validated code-editing model changes by reviewing GitHub pull requests and ensuring correctness through unit test verification. Developed prompt and multi-turn conversation handling requirements including prompt-injection scenarios with persona adherence and structured tool responses. • Fine-tuning for LLM instruction following (SFT/RLHF) • Evaluation via PR review and unit-test checks • Prompt engineering for tool-calling with JSON outputs • Multi-turn conversational behavior and red-teaming style prompt-injection tasks