At MDPI, I architect and deploy AI microservices for scholarly publishing, including enterprise NLP, computer vision, and LLM solutions. I established automated CI/CD and model lifecycle management pipelines using FastAPI and MLflow.
I built an AI English Editing Assistant by fine-tuning Mistral and Qwen for academic grammar correction, and established the MDPI academic GEC benchmark, where my SFT and DPO approach outperformed public LLM APIs. I also built a semantic search recommendation engine indexing academic profiles and publication records, improving reviewer recommendation acceptance by 30%.
