Researcher combining AI engineering with human communication. My research is on AI-augmented communication: the systems that generate, restructure and route language, and what they do to meaning, authorship and trust. I focus on health because that is where a change in the technology lands on someone's actual life. The aim is to explain and simplify health technology until people understand it, accept it, and benefit from it easily, in clinical work and in mental health alike.
I build the systems and I study what they do to the exchange. Computer science is the field; communication is the object.
My research is on AI-augmented communication: the systems that generate, restructure and route language once a model is in the loop. I come at it as a computer scientist. I build and evaluate the pipelines, and I study what they quietly change about the exchange in the process.
The questions I keep pulling on are about handover. When an agent compresses an encounter into a record, what survives and who decides? When a note is generated rather than written, where does accountability sit, and what does the audit trail actually prove? Communication was always the load-bearing layer of professional work. Agents make that visible by automating it.
Health is where the work has to land, because a change in the technology reaches someone's actual life: a patient reading a summary, a family deciding what to do, a clinician at the end of a shift. What I aim at is technology people understand and accept, in clinical work and in mental health alike, which means explaining it plainly and simplifying it rather than dressing it in jargon. A detail lost in a clinical summary is not a cosmetic defect.
Since 2025 I have worked as Content and Communications Specialist at Curvia, on research communication and student outreach across 25 countries. My path ran through digital media production in China, a BSc in Computer Science and Software Engineering at NUIST, and an MSc in Computer Science at Nanjing University of Posts and Telecommunications.
“The interesting failures are rarely in the model. They are in the handover.”
Each thread starts where language is doing load-bearing work in a professional setting, and an agent has just been invited into it.
What an agent keeps and loses when it summarizes, restructures or rewrites a human exchange. Comparing human-produced and model-produced versions of the same record, and locating exactly where meaning shifts.
Who is accountable when the writing is delegated. Provenance signals, edit history, and what a reader actually needs in order to trust a generated document.
Whether one finding can survive the trip between clinician, patient, student and machine. Health communication is the high-stakes case; research communication is the everyday one.
Three manuscripts, all currently under review.
Three manuscripts are in review. Titles and abstracts are shared on request while the process is running, and each row becomes a full entry once a decision lands.
Digital media to software engineering to the language layer of technical systems.
Research communication strategy for a global community of students and mentors across 25 countries. Editing, content systems, and student-facing programme communication.
Coursework and project work around AI-augmented communication, agent-mediated documentation, and health communication.
Software engineering foundations, with an earlier start in digital media production in China.
The beginning of the interest in communication as an engineered artifact: how a message is built, compressed, and received.
I am glad to talk about language model evaluation, agent-mediated writing and provenance, or the engineering side of health communication tools. The fastest route is email.