FM- Sequential Decision Making & Efficient Goal Driven Behavior Intern
IBM
This is for a 2025 summer internship with the following start dates: May - August or June - September for quarter system schools.
We are broadly interested in making foundation models (FMs) effective for sequential decision making in a range of tasks, particularly those related to efficient agents and their goal driven behavior. We are looking for interns with the following areas of interest:
Planning/action model elicitation, search components code elicitation (e.g., search guidance or search pruning components), solutions or models validation/verification, reasoning about action and change, planning for data.
Tasks of interest include:Using foundation models for deriving implicit information from textual description.Using foundation models to improve planning: before/during/after search.Using foundation models for validation/verification of solutions or planning models.Reasoning about action and change: efficient goal driven behavior. Pinpoint and tackle the difficulties in producing correct behavior efficiently.Planning for query decomposition, thoughts generation and evaluation.
We are broadly interested in making foundation models (FMs) effective for sequential decision making in a range of tasks, particularly those related to efficient agents and their goal driven behavior. We are looking for interns with the following areas of interest:
Planning/action model elicitation, search components code elicitation (e.g., search guidance or search pruning components), solutions or models validation/verification, reasoning about action and change, planning for data.
Tasks of interest include:Using foundation models for deriving implicit information from textual description.Using foundation models to improve planning: before/during/after search.Using foundation models for validation/verification of solutions or planning models.Reasoning about action and change: efficient goal driven behavior. Pinpoint and tackle the difficulties in producing correct behavior efficiently.Planning for query decomposition, thoughts generation and evaluation.
Confirm your E-mail: Send Email