Join us Partner with us

Publications

Everything we've written, from published articles to papers we're still working on. Each one has its own page with the abstract, the full list of authors and a BibTeX citation you can copy.

Published 7

  • Improving Human Sequential Decision Making with Reinforcement LearningHamsa Bastani, Osbert Bastani, Wichinpong Park SinchaisriAn algorithm learns from logs of people's multi-step decisions and turns the biggest gap between their play and an optimal policy into one short tip. In virtual-kitchen experiments the tips raised performance, and people combined them with their own strategies.
    Management Science, 72(1), 733–755, 2026Algorithmic Advice
  • Learning on the Go: Understanding How Gig Economy Workers Learn with Recommendation AlgorithmsShunan Jiang, Wichinpong Park SinchaisriDraws on more than a million delivery orders to show couriers get faster and more punctual with experience. Newcomers tend to follow the app's job recommendations, while seasoned workers increasingly go their own way.
    Proceedings of the ACM on Human-Computer Interaction, 9(7), 1–35, 2025Learning on the JobPlatform Work
  • Making ChatGPT Work for MeSamantha Keppler, Wichinpong Park Sinchaisri, Clare SnyderObserves 24 K-12 teachers using ChatGPT on their own work and sorts their requests into a handful of distinct ways of asking for help; most conversations stay within one of them.
    Proceedings of the ACM on Human-Computer Interaction, 9(2), 1–23, 2025AI at Work
  • The Impact of Behavioral and Economic Drivers on Gig Economy WorkersGad Allon, Maxime C. Cohen, Wichinpong Park SinchaisriUses ride-hailing data to model when on-demand drivers start and stop working. Pay matters, but so do habits such as quitting after reaching an earnings goal and staying on once a shift has momentum; incentives that account for both can add capacity without extra cost.
    Manufacturing & Service Operations Management, 25(4), 1376–1393, 2023Platform Work
  • Community Vibrancy and Its Relationship with Safety in PhiladelphiaWichinpong Park Sinchaisri, Shane T. JensenTurns Philadelphia block-party permits into two measures of neighborhood 'vibrancy' and links them to crime: more parties overall go with more crime, but a larger share of spontaneous parties goes with lower and falling crime.
    PLOS ONE, 16(12), e0257530, 2021
  • Managing the Gig Economy via Behavioral and Operational LensesWichinpong Park SinchaisriPark Sinchaisri's doctoral dissertation at the University of Pennsylvania; the gig-driver paper is one of its chapters.
    Ph.D. dissertation, University of Pennsylvania, 2021Platform Work
  • Pricing with Quality Perception: Theory and ExperimentWichinpong Park SinchaisriPark Sinchaisri's MIT master's thesis on pricing when buyers judge quality from price, combining a model with an experiment. It later grew into the markdown-pricing working paper.
    S.M. thesis, Massachusetts Institute of Technology, 2016

Under review 6

  • Managing Multihoming Workers in the Gig EconomyGad Allon, Maxime C. Cohen, Ken Moon, Wichinpong Park SinchaisriCombines one ride-hailing platform's data with public New York City trip data to estimate how drivers split time between competing apps. Guaranteed hourly pay can raise supply at fixed cost, streak bonuses curb switching, and wage floors need limits on congestion.
    Revise and resubmit, SSRN working paper 4502968, 2026Platform Work
  • Precise or Broad? The Reward-Learning Frontier in Algorithmic Advice DesignPhilippe Blaettchen, Wichinpong Park SinchaisriCompares two ways of turning an optimal policy into advice in EV-charging route experiments: an exact action versus a broader rule of thumb. Exact advice performs best while it is visible; broader advice can leave people better able to act alone in similar settings once advice is removed.
    Major revision, 2026Algorithmic AdviceLearning on the Job
  • Learning to Gig: Workforce Development under Worker AgencyHaonan Deng, Shunan Jiang, Xinpeng Qu, Wichinpong Park SinchaisriAsks how a platform should offer jobs when workers may decline and improve at different, unknown rates. Proposes a tractable benchmark and learning algorithms, including a variant that caps the risk of market failure, and tests them on delivery-platform data.
    Major revision, SSRN working paper 5387320, 2025Learning on the JobPlatform Work
  • Markdown Pricing with Quality PerceptionRim Hariss, Georgia Perakis, Wichinpong Park Sinchaisri, Yanchong (Karen) ZhengSome shoppers read a higher price as a sign of higher quality and are let down by smaller-than-expected discounts. The paper builds this into a retailer's markdown plan, calibrated with a consumer study, and finds that announcing markdowns in advance can pay off.
    Major revision, SSRN working paper 3634129, 2020
  • Act, Ask, or Defer? Authority Design for Agentic AI Service WorkflowsHaonan Deng, Wichinpong Park Sinchaisri
    SubmittedAI at WorkAlgorithmic Advice
  • Mechanism-Guided Generative AI Simulation for Pricing OperationsXinwei Chen, Xiexin Liu, Wichinpong Park Sinchaisri, Yixiang Xu
    SubmittedAI at Work

Working papers and preprints 5

In progress 7