Taimoor Ijaz

AI Model Evaluator (C/C++)

Turing Intelligence • Temps plein

Sep 2025 - Jan 2026 • 4 mos

Design, develop, and maintain efficient, high-quality C++ code to train and optimize AI models. Conduct evaluations (Evals) to benchmark model performance and analyze results for continuous improvement. Evaluate and rank AI model responses to user queries across diverse domains, ensuring alignment with predefined criteria. Develop comprehensive explanations and rationales for evaluations, showcasing excellent reasoning and technical expertise. Lead efforts in Supervised Fine-Tuning (SFT), including creating and maintaining high-quality, task-specific datasets. Collaborate with researchers and annotators to execute Reinforcement Learning with Human Feedback (RLHF) and refine reward models. Design innovative evaluation strategies and processes to improve the model's alignment with user needs and ethical guidelines. Create and refine optimal responses to improve AI performance, emphasizing clarity, relevance, and technical accuracy. Conduct thorough peer reviews of code and documentation, providing constructive feedback and identifying areas for improvement. Collaborate with cross-functional teams to improve model performance and contribute to product enhancements. Continuously explore and integrate new tools, techniques, and methodologies to enhance AI training processes.