CV
Contact Information
| Name | Sabrina Sadiekh |
| sadsobr7@gmail.com | |
| Website | https://sadsabrina.github.io/sabrina-sadiekh/ |
Experience
-
2026 - present Remote
R&D Research Lead
HiveTrace
- Architected proposal-driven lab structure spanning three tracks — XAI in Security, AI Security, and AI Safety.
- Supervise ~12 student research projects from problem formulation through publication.
- Co-authored 2 papers accepted to AINL 2026; 3 papers under ARR double-blind review.
-
2022 - present Remote
Independent Researcher
AikyamLab (with Chirag Agarwal, UVA)
- Research on mechanistic interpretability and AI safety in collaboration with an international group.
- Co-authored papers accepted at AAAI 2026 and ACL 2026.
-
2024 - present Lecturer & Program Creator
HSE University
- Courses: Advanced ML Topics, Mathematics, Explainable AI.
- Designed and delivered lectures and seminars for M.Sc. students.
-
2025 - 2025 Hybrid
AI Researcher
White Circle
- Analyzed and benchmarked proprietary guardrail models, evaluating robustness and bias dimensions.
-
2023 - 2024 Remote
Course Developer & Teaching Assistant
AI Education (EdTech)
- Authored materials in Machine Learning and Audio Deep Learning.
Education
-
2023 - present M.Sc.
Higher School of Economics (HSE University)
Artificial Intelligence
- Research focus on mechanistic interpretability and AI safety.
-
2019 - 2023 Petrozavodsk, Russia
B.Sc.
Petrozavodsk State University
Mathematics
- Thesis: Analysis of Explainable AI Approaches
- Developed taxonomy of XAI methods across data modalities.
- Formalized mathematical frameworks for explanation of deep models.
Publications
-
2026 GLiNER Guard: Unified Encoder Family for Production LLM Safety and Privacy
arXiv 2026
-
2026 Cross-Lingual Jailbreak Detection via Semantic Codebooks
arXiv 2026
Skills
Research (Expert): Explainable AI, Mechanistic Interpretability, Sparse Autoencoders, Linear Representation Hypothesis, AI Safety
Machine Learning (Expert): PyTorch, Transformers, Probing, Representation Analysis
Programming (Expert): Python, Julia
Languages
Russian : Native
English : Professional working proficiency
Interests
Research: Representation geometry, Platonic representation hypothesis, LLM internals, AI robustness
Education: Open-source XAI course (600+ students), Telegram blog (1300+ subscribers)