About me

Hi, I’m Karthik. I’m a researcher at the University of Pennsylvania building human-in-the-loop AI for evidence synthesis in medicine, with the Center for Integrative Global Oral Health and Penn NLP. I work with Chris Callison-Burch, Alonso Carrasco-Labra and Michael Glick, and with Sharath Chandra Guntuku on speech and health.

Before Penn: an M.S. in Computer Science at Syracuse, an internship at Siemens Healthineers, and software engineering at Broadridge in Hyderabad.

My research asks what it takes to put a language model inside a process that already has rules. EviStreams, the platform I lead, keeps domain experts in control of AI-assisted extraction for systematic reviews, and appears at EMNLP 2026. Across four clinical corpora and three model families, extraction quality depended far more on how a field was specified than on which model ran it. Separately, across 1,465 recorded interviews, combining what people say with how they say it predicted loneliness better than either signal alone.

I’m applying to PhD programs this cycle, in NLP and health AI.

Outside research: gym, travel, films, Indian food, photography. Reach out if you’d like to talk about evidence synthesis, health NLP, or PhD applications.

Recent News

• [Sep 2026] newsThe ADA released its new living clinical practice guideline on oral cancer, led by Penn Dental Medicine’s Center for Integrative Global Oral Health. Penn Dental News
• [Sep 2026] preprintPredictors of loneliness in older adults posted to arXiv.
• [Aug 2026] acceptedEviStreams accepted at EMNLP 2026, System Demonstrations.
• [May 2026] acceptedLoneliness work accepted to the 2026 Voice AI Symposium, hosted by the Bridge2AI-Voice Consortium in St. Petersburg, FL.
• [Mar 2026] posterPresented AI-Enhanced Living Evidence Synthesis in Oral Health at IADR 2026 in San Diego.
• [Jun 2024] joinedStarted at the University of Pennsylvania, building AI systems for clinical evidence synthesis.