My name is Yihan Jiang, ECE Ph.D. graduate from the University of Washington, Seattle (2016--2021).
Current Position: Research Scientist in Sentient.Foudation, remote, (2024-2025)
Previous Position: Founding Research Engineer/Manager in Aira Technology, Saratoga, CA (2020--2024)
Advisor: Professor Sreeram Kannan
PhD Research Area: Channel Coding, Meta Learning, and Federated Learning.
Current Interest: LLM, VLM, and agents.
My CV, Google Scholar, Github.
My Email: yij021 at uw dot edu
Very fortunate to work with my long-time collaborators: Professor Hyeji Kim, Professor Sewoong Oh, Professor Himanshu Asnani, and Professor Pramod Viswanath.
Please check our research blog on inventing channel coding via AI: https://deepcomm.github.io/
VLM, LLM and LLM agents, the next Internet.
(Feb 2025) Dobby-mini, a model that is strong at tone is launched. You might be offended by it, be open minded.
(Feb 2025) Open Deep Research, an agentic search framework, launched! Paper is out. Feel free to play with it on Sentient.chat!
(Nov 2024) Werewolf Agent competition launched! Have fun to build LLM agent that kills at playing Werewolf (Mafia).
(Aug 2024) Joined Sentient Foundation as research scientist/engineer.
---------------------------------------------------------------------
(March 2024) Demo RANGPT (now named Naavik) on MWC 24.
(Apr 10 2022) I have the pleasure of giving a tutorial on Physical Layer Communications via Deep Learning at WCNC 2022, with Professor Hyeji Kim. Check the slides.
(Mar 15 2021) Defend my thesis! Check my thesis slides.
(May 22 2020) New paper on SPAWC 2020. Joint channel coding and modulation via Turbo Autoencoder.
(May 4-8, 2020) New paper on ICASSP 2020. Feedback Turbo Autoencoder.
(Mar-2-2020) Joining Aira Technology, Saratoga, CA. Stealth mode.
---------------------------------------------------------------------
(Fed-8-2020) Turning 30! Still feel very curious about everything around, like a baby!
(Fed 2-7, 2020) To attend ITA 2020, to present Turbo Autoencoder: Deep learning based channel codes for point-to-point communication channels.
(Jan 27, 2020) PhD General Exam passed! Call me PhD candidate! slides.
(Dec 13, 2019) To present Improving Federated Learning Personalization via Model Agnostic Meta Learning in NeurIPS FL workshop 2019, Vancouver.
(Dec 12, 2019) To present Turbo Autoencoder: Deep learning based channel codes for point-to-point communication channels in NeurIPS 2019, Vancouver.
(Sep 25, 2019) Present an ongoing project of image embedding in Allerton!
(Sep 22, 2019) Finished a great summer internship with host Jakub Konečný, and co-host Keith Rush, in Google Beijing. Work on linking Federated Learning and Model Agnostic Meta Learning (MAML), with workshop paper accepted in NeurIPS FL workshop 2019, Vancouver.