CV
Academic and professional experience in computational physics, machine learning, and materials science.
General Information
| Full Name | Saba Kharabadze, PhD |
| Location | Pisa, Tuscany, Italy |
| Languages | English (native/bilingual), Georgian (native/bilingual), Russian (proficient) |
Education
-
2024 PhD in Physics
Binghamton University, State University of New York - Dissertation: Machine Learning and ab initio insights into the design of lithium-based materials
- Developed neural network potentials for materials discovery using PyTorch and C
- Published first-author research in npj Computational Materials (Nature Partner Journal) and PCCP
-
2017 Bachelor of Science in Physics
Free University of Tbilisi, Georgia - Minor in Computer Science and Mathematics
- Merit-based tuition scholarship recipient
Research & AI Engineering
-
2025 - present ML Research Engineer (Research Scientist)
CNIT RaSS Lab, Pisa, Italy - Winner of the IEEE Radar Challenge: Developed a Transformer-based pipeline for heart rate estimation, achieving a 65% improvement over the baseline
- Utilizing ML approaches for radar signal processing and target detection
- Building simulation frameworks for coverage analysis of mono- and multistatic radar systems
-
2025 - present Physics AI Model Validation Expert
Handshake MOVE Program (Remote) - Designing complex reasoning problems to stress-test AI models’ understanding of physics and logic
- Evaluating model responses and providing structured pedagogical feedback to improve reasoning capabilities
-
2018 - 2024 PhD Candidate (Research Engineering)
Binghamton University, Binghamton, NY - ML-Accelerated Discovery & High-Throughput Pipelines
- Built end-to-end pipelines scaling evolutionary search + ML potentials to 1M+ structures and 1TB of data
- Managed large-scale HPC campaigns with reproducible run tooling, logging, and structured outputs
- Identified 8 new stable structures redefining Li-Sn convex hulls
- Systems & Software Development
- Implemented an NPT barostat module in C for the MAISE software package; optimized performance-critical paths
- Administered two university-wide HPC Linux clusters (1,000+ cores); configured GPU nodes and job schedulers
- {“Installed and maintained the full scientific stack”=>”CUDA, PyTorch, JAX, VASP, and Intel/GNU compilers”}
- ML-Accelerated Discovery & High-Throughput Pipelines
Professional Experience
-
2024 - present Founding Member, Data Science / ML
Keepers - https://kprs.pro, Remote - Build ETL pipelines to ingest exchange data using Python, SQLAlchemy, and SQL; implement cleaning in pandas
- Develop anomaly detection and trend prediction on financial time-series using XGBoost and feature engineering
-
2017 - 2018 Back-end Developer
Eleven Wireless, Portland, OR (remote) - Developed Python middle layer interfacing SQL databases and ElasticSearch
- Deployed and maintained reporting services on AWS Linux instances with live patching
Teaching & Educational Technology
-
2015 - 2016 Co-Founder (Conceptual Design & Content)
Nebula, Tbilisi, Georgia - Co-founded an online platform providing Khan Academy-style video lectures and interactive exercises for national entrance exams
- Designed the technical architecture for the content delivery system and drafted specifications for automated student assessment
- Conceptualized logic for automated difficulty progression and structured data schemas for math/reasoning evaluations
- Authored and produced video lectures covering core STEM topics for the Georgian national curriculum
-
2018 - 2020 Head Teaching Assistant - Physics
Binghamton University, Binghamton, NY - Led lab sessions for introductory physics; earned 85% positive feedback
- Mentored and onboarded new TAs; created supplementary solution guides and review worksheets
-
2012 - 2018 Physics Olympiad Coach
42nd School of Physics and Mathematics, Tbilisi, Georgia - Coached students for international competitions; led the team to the 2016 International Zhautykov Olympiad
- Team awarded silver medals in both physics and mathematics
Skills
-
Technical Skills
- ML & Systems: JAX, PyTorch, Transformers, Distributed Training
- Programming: Python (pandas, NumPy, SciPy), C/C++, SQL, Bash
- Infrastructure: HPC (GPU nodes, Slurm/Schedulers), CUDA, Docker, Linux SysAdmin, AWS
- Evaluation: AI Model Validation, Reasoning benchmarks, Learning Science, Scaffolding
Honors and Awards
-
2025 - Winner, IEEE Radar Challenge
- Handshake AI Fellow, Handshake MOVE Program
-
2012 - Bronze Medal, 43rd International Physics Olympiad (Tallinn, Estonia)
- Bronze Medal, 8th International Zhautykov Olympiad (Almaty, Kazakhstan)