Postdoctoral Position In Mobility Choice Inference
The post-doctoral researcher will contribute to the following activities:
• Inference of mobility choice from heterogeneous evidence: transform unstructured behavioral information (free text from surveys, user feedback, incident/event messages, service alerts) into semantic factors (reliability, congestion/crowding, comfort, safety, cost, accessibility), then merge with structured traces and urban context (pricing, disruptions, events, exogenous signals) to infer probabilistic distributions of modal choices and mobility decisions.
• LLM-guided agents for complete travel chains: design LLM agents generating complete daily chains (activities, places/POI, times, durations, multi-modal transfers) guided by learned distributions and anchored by external tools (routing/POI/context), ensuring individual plausibility and aggregate statistical fidelity.
• Context adaptation and robustness: enable adjustment of mobility sequences under changing conditions (disruptions, weather, events, congestion), while maintaining diversity, realism, and reproducibility.
• Experimentation and evaluation: conduct experimental campaigns on France-Canada data (subject to agreements and governance rules), including Paris/Lyon (smart card, loop counts, OD/presence mobile aggregates, GPS survey, bike flows, text events, census/OD) and Montreal/Ottawa (mobile aggregates, smart card, planning requests, bikes, OD survey), with micro (chain coherence, temporal feasibility/transfers) and macro (OD, modal shares, temporal profiles) evaluations, as well as inter-city and inter-country transfer tests.
• Valorization: publications, scientific communications, contribution to project deliverables and open science dissemination.
Special mobility possibility at McGill (University of Montreal).
Ideal Profile
Required profile:
Doctorate (obtained) in computer science, transport engineering, applied mathematics/statistics, or related field.
Strong background in machine learning (probabilistic modeling and/or representation; generative models preferred).
Experience in LLMs/NLP and/or agent-based systems highly desired.
Very good skills in Python and ability to build reproducible experimental pipelines.
Ability to conduct independent research and write scientific articles in English.
Expected qualities: autonomy, rigor, team spirit, clear communication, respect for ethics and good practices (sensitive data).
Others:
Required skills: machine learning (probabilistic modeling, representation, generative models), LLMs/NLP and/or agent-based systems, fusion of multi-source data, advanced Python programming, reproducible experimentation (versioning, pipelines, documentation), scientific writing in English.
Education and professional experience: Doctorate in a relevant field (computer science, transport, applied statistics/mathematics). Experience with mobility data (smart card, aggregated OD, GPS), mobility simulation/agents (e.g., MATSim/SUMO) or integration of heterogeneous data is appreciated.
Environment, work context, hierarchical reporting: GRETTIA Laboratory (COSYS department) – multidisciplinary environment; close interaction with the supervisory team (Univ Eiffel and McGill); participation in scientific life (seminars, conferences); access to computing resources and datasets, in compliance with governance and confidentiality rules.
Knowledge: solid foundations in statistical and probabilistic learning; notions of generative models (VAE/GAN/diffusion); basics in NLP/LLMs (embeddings, prompting, agents/tools); principles of transport system modeling; notions of data privacy/ethics.
Know-how: designing models and architectures (LLM + probabilistic/generative models), setting up data preparation/fusion pipelines, implementing and evaluating generative agents (travel chains, intermodality, temporal constraints), conducting multi-city experiments, analyzing and presenting results, producing documented, traceable, and reproducible code, publishing and communicating.
Know-how: autonomy, rigor, sense of initiative, scientific curiosity, team spirit, clear communication, respect for ethics and good research practices; ability to work in an international context (possible mobility to McGill).