Summary
Moyan M is a Research Scientist with 11 years of applied AI and ML experience, currently building state-of-the-art generative AI systems for Instagram at Meta. He has led end-to-end AI products—from founding NLP toolkits and mobile-friendly distilled models to architecting long-context LLMs, multi-agent legal AI, and RAG systems for enterprise clients like Lenovo, Pepsi, and Microsoft. His work blends deep research (DPO, RLHF, MoE/dense LLMs, contrastive and self-supervised learning) with pragmatic ML-Ops, model compression, and cost-aware deployment to deliver production-ready, scalable solutions. Notably, he devised knowledge distillation frameworks (SEAD/ESEAD) achieving large model speedups and compression while preserving performance, and built a multi-LoRA conversational engine that automates business analysis. Comfortable across multimodal, multilingual, and long-context problems, he also has strong product instincts from director-level roles and startup founding experience. Outside work he channels creative curiosity into photography and experimental coffee brewing—habits that mirror his iterative, design-driven approach to AI.
11 years of coding experience