UW PhD Shangbin Feng enters 2026-27 faculty market with participatory AI research agenda
shangbinfeng · x · 2026-10-06
Shangbin Feng, final-year PhD at University of Washington (advisor Yulia Tsvetkov, former NVIDIA/Google intern), is on the 2026-2027 faculty job market and will be at COLM. He proposes "participatory AI": beyond closed AI (few companies decide data and values) and open-weight AI, modular systems built bottom-up by diverse stakeholders who keep ownership of contributions. Research threads include:
- Training models to complement systems they join (Knowledge Card, Modular Pluralism)
- Alignment via collaborative feedback (Sparta Alignment, Stackelberg Alignment)
- Task adaptation through evolutionary swarm search over weights and orchestrations (Model Swarms, Data Swarms)
- Scaling participatory deployment with safeguards against malicious contributions, treating participation as a new scaling axis (Scaling Participation, Among Us)
Also interested in knowledge/factuality, networks, pluralism, multi-agent, and safety.
More from Companies & People
- COLM 2026 Kicks Off in San Francisco as Public Registration Sells Out — stanfordnlp · 2026-10-07
- Jon Stewart Roasts Meta's Muse AI Agent; Alexandr Wang Fires Back on X — alexandr_wang · 2026-10-07
- Agent PRs up 9x in 8 months on GitHub: the 2026 tech industry snapshot — The Pragmatic Engineer · 2026-10-07
- Artist Alexander Reben joins Google as Artist in Residence after OpenAI and Meta stints — artBoffin · 2026-10-07
- Short practitioner posts beat polished case studies for showing domain expertise — brandon_galang · 2026-10-07
- Swami Sivasubramanian: Companies Stuck Using AI as Assistants, Industry Under 1% There — BradPorter_ · 2026-10-07