# DrawOffer: Evidence-Grounded Longitudinal User-State Modeling for Career Support

DevFeed: [DrawOffer: Evidence-Grounded Longitudinal User-State Modeling for Career Support](<https://devfeed.tech/articles/drawoffer-evidence-grounded-longitudinal-user-state-modeling-for-career-support-115376.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.02793>)

Author: Lele Xu, Yuning Bian, Yifan Yuan

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Tech Careers](<https://devfeed.tech/topics/tech-careers.md>), [Interactive AI Systems](<https://devfeed.tech/topics/interactive-ai-systems.md>)

Tags: [ablation](<https://devfeed.tech/tags/ablation.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [artifacts](<https://devfeed.tech/tags/artifacts.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [cache](<https://devfeed.tech/tags/cache.md>), [career](<https://devfeed.tech/tags/career.md>), [checkpoint](<https://devfeed.tech/tags/checkpoint.md>), [comparison](<https://devfeed.tech/tags/comparison.md>), [confidence](<https://devfeed.tech/tags/confidence.md>), [context](<https://devfeed.tech/tags/context.md>), [contribution](<https://devfeed.tech/tags/contribution.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [developer](<https://devfeed.tech/tags/developer.md>), [development](<https://devfeed.tech/tags/development.md>), [experience](<https://devfeed.tech/tags/experience.md>), [full](<https://devfeed.tech/tags/full.md>), [modeling](<https://devfeed.tech/tags/modeling.md>), [state](<https://devfeed.tech/tags/state.md>), [support](<https://devfeed.tech/tags/support.md>), [user](<https://devfeed.tech/tags/user.md>)

## AI overview

DrawOffer separates language-model extraction of observations from deterministic inference over source-linked records to track uncertain abilities, experience assessments, and revisable issues, withholding judgments when evidence is insufficient. On 37 test personas it passed 169 of 210 checkpoints; checkpoint-defined issue precision was 0.875 and recall 0.636. A cache-only development ablation that replaced answer-evidence requirements with confidence-only settlement reduced majority-vote passes from 306/370 to 293/370 and issue precision from 0.927 to 0.725. Results are exploratory: developer-authored scenarios, unmatched baseline models, historical test exposure, and incomplete public artifacts limit generalization and independent reproduction.

## Source excerpt

arXiv:2611.02793v1 Announce Type: new Abstract: Career-support assistants must distinguish user claims from evidence of current ability and withhold judgments when evidence is insufficient. DrawOffer separates language-model observation extraction from deterministic inference over source-linked records to maintain uncertain ability estimates, experience assessments, and revisable issues. We evaluate 102 fictional personas with 580 scored checkpoints, a historical full-context and Mem0 comparison, and three auxiliary knowledge-tracing datasets. On 37 test personas, DrawOffer passes 169/210 checkpoints; checkpoint-defined issue precision and recall are 0.875 and 0.636. Across three development simulations, 87.0% of checkpoint verdicts agree, including persistent failures. A cache-only development ablation replacing answer-evidence requirements with confidence-only settlement reduces majority-vote passes from 306 to 293 of 370 and issue precision from 0.927 to 0.725. Equalizing positive source weights improves predictive metrics without changing total passes; removing decay gives mixed results. These findings support an exploratory systems contribution centered on evidence sufficiency. Developer-authored scenarios, unmatched baseline models, historical test exposure, and incomplete public artifacts limit generalization and independent reproduction.