Thesis Worker 30 hp - Adaptive Hierarchical Inference with Foundation models

Job description** 1. Conduct a literature study on compact foundation models, hierarchical inference, uncertainty estimation and calibration, learning to defer, and adaptive resource aware edge-cloud offloading. 2. Select and deploy a compact open-source model on a representative edge or HPU platform. Characterize model quality, latency, throughput, memory footprint, model size, and, where feasible, energy consumption. 3. Implement and evaluate a confidence or uncertainty-estimation method for the local model. 4. Develop an adaptive threshold-learning or cloud-escalation strategy that jointly considers uncertainty, task quality, and other relevant factors such as latency and operational cost. 5. Compare local-only, remote-only, static-threshold, adaptive-routing, and retrospective-oracle baselines. Report repeated trials and uncertainty for task quality, calibration, offload rate, latency, communication, memory, and energy where feasible. 6. Deliver reproducible code, experiment configurations, documented datasets and models, analysis scripts, and a clear account of limitations and deployment assumptions. An optional extension is a theoretical performance model, regret analysis, or a clearly scoped multimodal experiment. **Expected outcome** The expected result is a reproducible prototype and evaluation pipeline that combines an edge-deployable foundation model, calibrated uncertainty estimates, and an adaptive escalation policy. Subject to result quality and confidentiality constraints, the work may also contribute to a scientific publication. References \[1\] V. N. Moothedath, J. P. Champati, and J. Gross, "Getting the Best Out of Both Worlds: Algorithms for Hierarchical Inference at the Edge," IEEE Transactions on Machine Learning in Communications and Networking, 2024. \[2\] C.-H. Chang, A. P. Behera, S. Zhang Pettersson, and J. Gross, "A Cost-Aware Hierarchical Cascade for Anomaly Detection at the Edge in Connected Vehicles," ACM/IEEE Symposium on Edge Computing, 2025. \[3\] Z. Liu et al., "MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases," ICML 2024. [https://arxiv.org/abs/2402.14905](https://arxiv.org/abs/2402.14905) **Education/line/direction** Master's programmes in Data Science, Machine Learning, Computer Science, Embedded Systems, Electrical Engineering, Engineering Physics, Applied Mathematics, or a similar field. Strong Python skills and experience with a deep-learning framework are expected. Familiarity with foundation-model inference, uncertainty estimation or calibration, embedded or edge platforms, multimodal or time-series learning, or performance profiling is beneficial. | | | | --- | --- | | Number of students | 1 | | Start date for the thesis project | January 2027, or as agreed | | Estimated timescale | 20 weeks | | Location | Within the TRATON GROUP; exact location and hybrid arrangement to be agreed | **Contact person and supervisor** Sophia Zhang Pettersson, Senior Data Scientist sophia.zhang.pettersson@scania.com Juan Carlos Andresen, Unit Manager, Research Advisor juan-carlos.andresen@scania.com **Application** Your application should contain a CV, personal letter, and copies of grades. Application period and deadline: 2026-10-31, applicants may be assessed continuously until the position is filled. **A background check might be conducted for this position. We are conducting interviews continuously and may close the recruitment earlier than the date specified.** Requisition ID: 33387 Number of Openings: 1.0 Part-time / Full-time: Full-time Permanent / Temporary: Temporary Country / Region: SE Location(s): Södertälje, SE, 151 38 Required Travel: 0% Workplace: On-site [Apply now »](https://jobs.scania.com/talentcommunity/apply/1442711733/?locale=en_US)