Pareel Amre
Master Thesis @Scads.Ai Dresden Leipzig
Signup · Get unlimited contacts
WORK HISTORY
Master Thesis @Scads.Ai Dresden Leipzig
Dresden, DE
Investigated the rationale and decision processes of large language models (LLMs) for future event summarization, analyzing model attention, intermediate reasoning steps, and temporal inference chains.Designed experiments to probe how LLMs weigh historic event context, causal relations, and narrative cues when summarizing possible future scenarios.Developed and evaluated frameworks (e.g. attribution, chain-of-thought probing) to expose rationale traces and identify model biases, inconsistencies, and temporal reasoning limitations.Suggested enhancements (e.g. hybrid models, explicit temporal constraints, rationale regularization) to improve LLM summarization quality and interpretability for predictive summarization tasks.
EDUCATION
Technische Universität Dresden
Master of Science - MS
Manipal Institute of Technology
Bachelor of Technology - BTech
ABOUT PAREEL AMRE
I am a AI Researcher & Machine Learning Engineer specializing in computational modeling, simulation, and hybrid AI systems. My work centers on combining physics-based modeling with data-driven methods to build robust, interpretable models that generalize well across real-world systems. My competencies span: • Designing neural-physics hybrid models in Python / PyTorch / TensorFlow • Time-series and spatio-temporal prediction using RNNs / Transformers • Model validation using empirical datasets and cross-disciplinary collaboration • Data visualization, simulation pipelines & reproducible research I’m passionate about solving real-world problems in domains like engineering, materials, energy, or environmental systems by leveraging AI and simulation.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.