deric pang

work

  • llm & vlm post-training for design, design text and image embeddings (2023 - )

  • google

    llm distillation, search query understanding, entity relationship mining (2019 - 2023)

  • representation learning of syntax for llms (2018 - 2019)

  • reinforcement learning in a game engine (2018)

  • training autonomous robots in simulation (2018)

  • amazon

    deep learning framework for speech recognition (2017)

  • code generation, fault localization (2015 - 2018)

projects