Summary
Shutian Ma is an NLP and GenAI engineer with eight years of experience building production-ready language and multimodal systems, currently working on generative AI annotation for Meta through Tundra Technical Solutions. He has hands-on ML engineering experience from roles at Tencent and Axon, where he optimized audio-to-text pipelines, speaker diarization, and LLM summarization for challenging real-world data like bodycam recordings. Shutian has deep practical knowledge in RAG, dataset curation, and deployment—having parsed large Chinese PDFs, implemented LangChain retrieval modules, and containerized ComfyUI deployments for clients. His background blends academic rigor (doctoral studies and visiting scholar work) with product-focused engineering, including leading annotation teams and applying LoRA fine-tuning to domain chatbots. Based in San Jose, he contributes to Simplified Chinese LLM annotation efforts that supported a Meta AI publication—an uncommon bridge between hands-on labeling and published research. Outside work he emphasizes a balanced approach to life, encapsulated by his GitHub bio: "live happy live free."
8 years of coding experience
5 years of employment as a software developer
Doctor Informatics, Doctor Informatics at Nanjing University of Science and Technology
Chinese, English