PeopleofSuperintelligence

Researcher

Kartik Goyal

Kartik Goyal is a natural language processing researcher whose probabilistic models look for latent structure in language and in cultural artifacts such as early modern printed books. He studied computer science and physics at BITS Pilani, earned a master's degree in 2014 and a PhD in 2021 at Carnegie Mellon's Language Technologies Institute, and worked on machine translation at Safaba Translation Solutions before the doctorate. His papers include a 2018 AAAI method that trains sequence models through a continuous relaxation of beam search, a 2020 generative model for the typography of early modern printing, and the 2024 ACL paper MAP's not dead yet, which received an Outstanding Paper Award. From 2021 to 2023 he was a research assistant professor at the Toyota Technological Institute at Chicago, and in 2023 he joined Georgia Tech. At Georgia Tech he is an assistant professor in the School of Interactive Computing.

Timeline

  1. 2012–2014

    BITS Pilani and a CMU master's

    He earns a B.E. (Hons.) in computer science and an M.Sc. (Hons.) in physics at the Goa campus of BITS Pilani in 2012. In 2014, he completes a master's degree at Carnegie Mellon's Language Technologies Institute, working on distributional semantics with Eduard Hovy. Before starting his PhD, he works as a machine translation researcher at Safaba Translation Solutions.

  2. 2017–2018

    Continuous relaxation of beam search

    In 2017, he coauthors A Continuous Relaxation of Beam Search for End-to-end Training of Neural Sequence Models with Graham Neubig, Chris Dyer and Taylor Berg-Kirkpatrick; it appears at AAAI in 2018. The paper replaces beam search with a continuous approximation so that a model trains on the loss of the final decoded output, and it improves results on named entity recognition and CCG supertagging.

  3. 2020

    Typographical analysis of early printing

    His ACL 2020 paper with Chris Dyer, Christopher Warren, Maxwell G’Sell and Taylor Berg-Kirkpatrick proposes a probabilistic generative model that clusters glyph images from early modern printed documents into typeface templates. The neural editor model beats the Ocular clustering baseline and a VAE at unsupervised discovery of typefaces in mixed-font documents. It analyzes glyph shapes in early modern books such as Leviathan by Thomas Hobbes, published in 1651.

  4. 2021–2023

    PhD and TTIC faculty post

    He earns his PhD in 2021 from Carnegie Mellon's Language Technologies Institute, jointly advised by Chris Dyer and Taylor Berg-Kirkpatrick. He serves as a research assistant professor at the Toyota Technological Institute at Chicago until 2023.

  5. 2022

    Energy language models

    In 2022, he coauthors Mix and Match: Learning-free Controllable Text Generation using Energy Language Models with Fatemehsadat Mireshghallah and Taylor Berg-Kirkpatrick for ACL. At TTIC, he is course instructor for Advanced Natural Language Processing (TTIC-31210).

  6. 2023

    Joining Georgia Tech

    He joins the School of Interactive Computing at Georgia Tech as an assistant professor in 2023, starting his first semester there. He aims to build a new lab with collaborators and students to develop statistical tools for the analysis and processing of natural language.

  7. 2024

    An ACL Outstanding Paper Award

    His paper MAP’s not dead yet: Uncovering true language model modes by conditioning away degeneracy, written with Davis Yoshida and Kevin Gimpel, receives an Outstanding Paper Award and the SAC Award for Generation at ACL 2024. It shows that the length-conditional modes of machine translation models and language models are more fluent and topical than their unconditional modes.

  8. As of October 2026

    Assistant professor at Georgia Tech

    He is an assistant professor in the School of Interactive Computing at Georgia Tech, working on natural language processing and machine learning problems that involve probabilistic models of latent structure in naturally occurring data. In 2025, he coauthors Train It and Forget It: Merge Lists are Unnecessary for BPE Inference in Language Models with Tomohiro Sawada for EMNLP.

Career

  1. Since 2023Georgia Institute of TechnologyAssistant Professor
  2. 2021–2023Toyota Technological Institute at ChicagoResearch Assistant Professor
  3. Safaba Translation SolutionsMachine Translation Researcher

Education

  1. 2021Carnegie Mellon UniversityPhD
  2. 2014Carnegie Mellon UniversityMaster's degree
  3. 2012Birla Institute of Technology and Science, PilaniB.E. (Hons.) in Computer Science; M.Sc. (Hons.) in Physics

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Kartik Goyal. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Georgia Institute of Technology. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  3. People shaping AI work at a median of 3 organizations over a career: 3 for university starters and institute and lab starters, versus 2 for company startersIncludes Georgia Institute of Technology. Across 6,853 people shaping AI with a dated work role, we estimate a median of 3 parent organizations over a career (mean 3.6), comparing role groups, starting sectors and starting decades from the 1950s to 2020–2026.
  4. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorateIncludes Georgia Institute of Technology. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorate. Across 2,116 people with both an undergraduate degree and a doctorate recorded, we trace which undergraduate schools feed which doctoral programs, how often students stay at the same school, and how paths differ by birth-country region.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Kartik Goyal yet. Send us a free-licensed photo.

  1. kgoyal8 | College of Computing cc.gatech.edu. Checked 21 October 2026.
  2. Kartik Goyal (TTIC) – Appropriate Inductive Bias for Neural Probabilistic Modelling of Data – Department of Computer Science cs.uchicago.edu. Checked 21 October 2026.
  3. Kartikgo cs.cmu.edu. Checked 21 October 2026.
  4. [1708.00111] A Continuous Relaxation of Beam Search for End-to-end Training of Neural Sequence Models arxiv.org. Checked 21 October 2026.
  5. A Probabilistic Generative Model for Typographical Analysis of Early Modern Printing - ACL Anthology aclanthology.org. Checked 21 October 2026.
  6. Assistant Professor Explores Analytical and Practical Benefits of Statistical NLP | College of Computing cc.gatech.edu. Checked 21 October 2026.
  7. Kartikgo kartikgo.github.io. Checked 21 October 2026.
  8. MAP’s not dead yet: Uncovering true language model modes by conditioning away degeneracy - ACL Anthology aclanthology.org. Checked 21 October 2026.
  9. Kartik Goyal | School of Interactive Computing ic.gatech.edu. Checked 21 October 2026.