Saloni Tuli

I’m a highly analytical Multimodal AI Specialist and Prompt Engineer, bridging technical computer science fundamentals with advanced business management frameworks. I engineer high-fidelity datasets, multi-turn instructional logic, and behavioral training workflows to scale Large Language Models (LLMs), Large Action Models (LAMs), and autonomous AI agents.

At SoftAge Information Technology Limited, I’ve engineered and validated 30,000+ high-fidelity multimodal training datasets, mapping complex text prompts to screen-recorded behavioral actions for enterprise-grade LAMs. I maintain high-throughput performance—averaging 80–100 precise prompt recordings daily—and deliver up to 300 complex multi-turn interactions during high-volume production sprints.

I specialize in Reinforcement Learning from Human Feedback (RLHF), rigorous model benchmarking, and end-to-end data quality auditing. I direct sanity and regression evaluation workflows, debugging and evaluating 1,500+ conversational and action-based prompts to optimize instruction compliance and meet zero-error enterprise production milestones.