Top GEO Researchers

Aggarwal led the November 2023 preprint that named generative engine optimization and treated visibility inside generative answers as a black-box problem researchers could measure. KDD 2024 later accepted the work.

GEO-bench and the paper's measured lifts still give commercial teams a shared academic starting point when they need more than screenshots.

Profile: Pranjal Aggarwal on People

Narasimhan is the senior academic on the original GEO paper and a Princeton computer scientist whose earlier work includes OpenAI's first GPT paper and agent systems such as ReAct and Tree of Thoughts.

The systems marketers try to influence descend partly from research he helped create, which is why his name sits at both the naming of GEO and the stack upstream of agentic search.

Profile: Karthik Narasimhan on People

With Mahe Chen, Xiaoxuan Wang, and Kaiwen Chen, Koudas compared AI search with Google across verticals, languages, and phrasings. The study found AI engines lean heavily on earned media and favor large brands.

Practitioners cite the work when they need academic evidence that third-party corroboration belongs at the center of a GEO program.

Profile: Nick Koudas on People

Xiong's CMU group released AutoGEO, an ICLR 2026 framework that learns generative-engine preference rules and rewrites web content for more traction while tracking generative-engine utility.

AutoGEO moves the field from hand-crafted rewrite heuristics toward rule extraction, plug-and-play API rewriting, and a compact reinforcement-learning model that commercial teams can cite as a systems baseline.

Profile: Chenyan Xiong on People

With Aounon Kumar, Lakkaraju showed that a crafted strategic text sequence on a product page could push a language model to name that product its top recommendation, and tied the risk to fair market competition.

The study remains the clearest academic line between legitimate visibility work and manipulation that the commercial GEO field still argues over.

Profile: Himabindu Lakkaraju on People

Peng co-authored E-GEO, the e-commerce generative engine optimization testbed that pairs thousands of intent-rich shopping queries with Amazon listings and measures rank change after rewriting.

The work gives sellers and researchers a commercial ranking objective instead of only impression proxies from earlier web GEO papers.

Profile: Tianyi Peng on People

Farias is a senior author on E-GEO and a long-running MIT Sloan researcher on decision systems, experimentation, and large-scale ranking problems.

E-GEO's prompt meta-optimization and red-team defenses sit in the same lineage as his earlier work on making optimization claims measurable under real commercial constraints.

Profile: Vivek Farias on People

Murahari co-authored the KDD 2024 GEO paper and GEO-bench while doing Princeton PhD work on NLP systems that find and use information under real retrieval conditions.

Equal-contribution notes on the preprint place him beside Aggarwal on the measurement and benchmark design commercial teams still cite.

Profile: Vishvak Murahari on People

Deshpande co-authored GEO and GEO-bench while completing an NLP PhD at Princeton with Narasimhan, with wider work spanning AI safety, efficient architectures, and multilingual models.

He treats visibility inside generative engines as a measurable systems problem rather than a marketing slogan.

Profile: Ameet Deshpande on People

Kumar led Manipulating Large Language Models to Increase Product Visibility with Lakkaraju, releasing code that shows how strategic text sequences can steer product recommendations.

Coverage in major news outlets made the paper a reference for platforms and agencies that need a concrete example of recommendation-steering risk.

Profile: Aounon Kumar on People

Xu is corresponding author on FeatGEO, the ACL 2026 paper that reframes generative citation optimization as feature-level multi-objective control over structure, content, and language.

The work argues that token-level rewrite tricks miss levers that matter when the goal is citation inside generative answers.

Profile: Peilan Xu on People

Sharma co-authored Beyond the Blue Link, the Hypertext 2026 paper that formalizes answer engine optimization and introduces Generative Share of Voice for stochastic retrieval.

His contribution sits at the seam of hypertext structure and machine-mediated attribution, which is why the paper travels beyond marketing blogs.

Profile: Raghav Sharma on People

Kotwal co-authored Beyond the Blue Link and helped validate how authored structure relates to citation outcomes across domains under stochastic retrieval.

Generative Share of Voice gives AEO and GEO teams a probabilistic vocabulary when answers change across runs.

Profile: Adit Kotwal on People

Rajpurohit is an independent co-author of the founding GEO paper, listed beside Princeton and other lab contributors on the KDD author line.

His seat shows that the field's naming paper was not only a single-campus product, which still matters when agencies cite the full author list.

Profile: Tanmay Rajpurohit on People

Kalyan co-authored GEO and helped shape the evaluation design behind GEO-bench, including how visibility gains were compared across domains.

Public summaries of the work credit him with keeping measurement next to the optimization claim rather than treating either alone.

Profile: Ashwin Kalyan on People

Van Buskirk co-led, with Kevin Indig, an independent UX study of Google's AI Overviews that used think-aloud recordings to show how people read AI answers.

The method filled a gap click logs cannot: confusion, trust moments, and scanning patterns that marketers need when they brief AI Overview projects.

Profile: Eric Van Buskirk on People

Druck leads scientists and engineers building AI tools for organic growth at Graphite after a PhD at UMass Amherst with Andrew McCallum and research seats at Yahoo and Yummly.

His applied lab work turns academic NLP habits into measurement systems agencies use when generative visibility has to survive client reporting.

Profile: Gregory Druck, PhD on People

Blyskal leads AI strategy at Profound and publishes large-scale prompt and citation studies that practitioners use to see how engines choose sources.

His measurement work sits between academic GEO-bench habits and the commercial dashboards buyers hire, which is why conference audiences treat the datasets as research infrastructure.

Profile: Josh Blyskal on People

Petrovic publishes technical research on how generative search systems retrieve, attribute, and surface sources, including architecture write-ups that agencies use when they need more than a vendor pitch.

DEJAN's public methods treat AI-search visibility as an information-retrieval problem with measurable parts.

Profile: Dan Petrovic on People

Volpini's WordLift work treats entities and knowledge graphs as the substrate generative engines use when they decide whom to cite.

Applied research demos and conference sessions on AI Mode and entity-aware content give practitioners a lab-adjacent playbook for corroboration and structured identity.

Profile: Andrea Volpini on People