PeopleofSuperintelligence

Researcher

Hai Pham

Hai Pham is a machine-learning researcher whose work spans multimodal language models, text recognition, and AI evaluation. He studied computer science at Auburn, earned a master's in Language Technologies at Carnegie Mellon in 2018, and completed a PhD in Language and Information Technology in 2023. His early research explored multimodal sentiment, cross-modal representations, and handwriting recognition for noisy documents. At Reka, he contributed to Flash, the Core, Flash, and Edge report, and Vibe-Eval, an open benchmark for multimodal chat. As of October 2026, he works at Reka AI.

Now at
Reka AI
Works
18
Citations
278
h-index
9

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 2017–2022

    Internships at Boeing, Microsoft and AWS

    He works as a research intern at Boeing in 2017 and Microsoft in summer 2021 and 2022, then as an applied scientist intern at Amazon AWS in fall 2022.

  2. 2018

    Language technologies at CMU

    He graduates in computer science at Auburn before earning a master's in Language Technologies at Carnegie Mellon in 2018.

  3. 2018

    A multimodal sentiment model

    In 2018, he coauthors Seq2Seq2Sentiment, which learns joint multimodal representations with sequence-to-sequence methods for sentiment analysis.

  4. 2019

    Found in Translation

    In 2019, he coauthors Found in Translation, which learns robust joint representations by translating between modalities and supports sentiment prediction with missing or perturbed inputs.

  5. 2020

    Handwriting recognition from noisy logs

    In 2020, he coauthors a handwriting-recognition system that learns from limited, noisy maintenance logs through word segmentation and word recognition.

  6. 2023

    A representation-learning PhD

    In 2023, he completes a PhD in Language and Information Technology at Carnegie Mellon with a dissertation titled Towards Efficient and Scalable Representation Learning.

  7. February 2024–May 2024

    Reka models and Vibe-Eval

    In February 2024, Reka credits him on the Flash team behind its 21B multimodal and multilingual language model. In April 2024, he coauthors the Core, Flash, and Edge report, which describes models that process text, images, video, and audio. In May 2024, he coauthors Vibe-Eval, an open benchmark of 269 visual-understanding prompts for multimodal chat models.

  8. As of October 2026

    Work at Reka AI

    He works at Reka AI.

Sources and credits

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

  1. Hai Pham cs.cmu.edu. Checked 14 October 2026.
  2. Hai Pham - Language Technologies Institute - School of Computer Science - Carnegie Mellon University lti.cmu.edu. Checked 14 October 2026.
  3. [1807.03915] Seq2Seq2Sentiment: Multimodal Sequence to Sequence Models for Sentiment Analysis arxiv.org. Checked 14 October 2026.
  4. [1812.07809] Found in Translation: Learning Robust Joint Representations by Cyclic Translations Between Modalities arxiv.org. Checked 14 October 2026.
  5. [2008.08148] Robust Handwriting Recognition with Limited and Noisy Data arxiv.org. Checked 14 October 2026.
  6. Reka Flash: Efficient and Capable Multimodal Language Models reka.ai. Checked 14 October 2026.
  7. [2404.12387] Reka Core, Flash, and Edge: A Series of Powerful Multimodal Language Models arxiv.org. Checked 14 October 2026.
  8. [2405.02287] Vibe-Eval: A hard evaluation suite for measuring progress of multimodal language models arxiv.org. Checked 14 October 2026.
  9. David P. Woodruff cs.cmu.edu. Checked 14 October 2026.