PhD candidate in experimental physics in the Kläui Group at Johannes Gutenberg University Mainz. I make magnetic materials a few atomic layers thick, build devices out of them, and write the software that measures them.
"Aditya Kumar" is a common name, common enough that a conference once accused me of double-submitting abstracts because of a namesake. So, to be specific: this is the one who works on two-dimensional magnets and spintronics. The full record is at aditya-kumar.de.
Exchange bias in van der Waals antiferromagnet/ferromagnet heterostructures, tunnelling magnetoresistance in CrSBr, and magnon transport in Pt/CrPS4/Pt. The work runs end to end: mechanical exfoliation and dry-transfer stacking, lithography and etching, then cryogenic transport, SQUID magnetometry, NV magnetometry, Raman spectroscopy and SEM/EDX. Five peer-reviewed publications during the PhD, three as first or shared-first author, one as lead author of an invited perspective.
The group's cryogenic transport measurements run on software I wrote. It holds a measurement for 20 to 30 hours unattended, with the magnet and temperature safety limits enforced by a state machine in code rather than by whoever happens to be watching the screen. Swapping an instrument is a YAML edit, and every real driver ships with a simulated twin so the stack can be developed and tested with no hardware attached.
I started and lead the lab's effort to make its own knowledge usable by AI agents. A retrieval system over the group's 234-paper corpus answers questions with a citation that resolves to the exact supporting sentence in the exact paper. The same treatment is being applied to the internal procedures wiki, and to instrument control an agent can drive within the safety envelope. One strand runs end to end; the other two are at an early stage.
| Repository | What it is |
|---|---|
CryoSoft Python |
PyQt6 desktop application controlling cryostat measurement systems. Six layers, raw drivers up to GUI, each depending only on the one below. Safety-enforcing state machine, YAML station registry, HDF5 output with automatic metadata capture, GUI panels generated by introspecting the instrument wrappers. Built for the Kläui Lab. |
TracEx Python |
Extracts experimental measurements from condensed-matter papers and grades every value against the sentence it came from. Sentence IDs are stamped before any model runs, deterministic extraction goes first, then a single schema-constrained LLM call per paper, then a re-check of each cited sentence that assigns an A/B/C provenance grade. |
EB-in-FGT_CrSBr Jupyter |
Data and analysis behind the switchable exchange bias in Fe3GeTe2/CrSBr result. AHE transport, SQUID, Raman, AFM. Small 21, e06284 (2025). |
EB-in-FGT-MnPS3 Jupyter |
Data and analysis for thermal modulation of exchange bias in MnPS3/Fe3GeTe2. AHE, SQUID, NV magnetometry, Raman, VASP. Adv. Mater. 36, 2403685 (2024). |
EB-in-FGT-CrPS4 Jupyter |
Data and analysis for preset-field-induced exchange bias in Fe3GeTe2/CrPS4 with surface oxides. ACS Nano 18, 8383 (2024). |
Quantum-Search-Algorithm Mathematica |
Implementation of the quantum-random-walk search algorithm of Shenvi, Kempe and Whaley (2003). |
FMR-simulation Jupyter |
Ferromagnetic resonance simulations for Larmor frequency and spin relaxation time, written while building the VNA-FMR setup for my MSc thesis. |
Every paper of mine with an accompanying repository has its data and analysis code published alongside it, so the figures can be regenerated from the raw measurement files.
- Prospects and challenges for exchange bias in van der Waals heterostructures. Newton 2, 100469 (2026). Invited perspective, lead author. 10.1016/j.newton.2026.100469
- Switchable exchange bias from correlated domain structures in orthogonally coupled AF/FM van der Waals heterostructures. Small 21, e06284 (2025). First author. 10.1002/smll.202506284
- Identifying the origin of thermal modulation of exchange bias in MnPS3/Fe3GeTe2 van der Waals heterostructures. Adv. Mater. 36, 2403685 (2024). 10.1002/adma.202403685
- Harnessing van der Waals CrPS4 and surface oxides for nonmonotonic preset-field-induced exchange bias in Fe3GeTe2. ACS Nano 18, 8383 (2024). 10.1021/acsnano.3c13034
- Mechanisms of electrical switching of ultrathin CoO/Pt bilayers. Nano Lett. 24, 1471 (2024). 10.1021/acs.nanolett.3c02890
- Tailoring spin-to-charge conversion efficiency via microwave frequency in La0.67Sr0.33MnO3/Pt bilayers. SPIN 14, 2340019 (2024). 10.1142/S2010324723400192
Preprint. Large bias-tunable magnetoresistance from spin-dependent interlayer hybridization in van der Waals antiferromagnet CrSBr-based heterostructures. arXiv:2608.11389. Shared first author. A junction with no magnetic electrodes at all, reaching 350% magnetoresistance at 20 K, where the antiferromagnetic order of the CrSBr barrier itself sets the barrier height and the bias voltage sets the size of the effect. Not peer-reviewed yet, so it has no DOI. The mechanism and the figures are on my site.
The full list, with citations, is on Google Scholar.
Fabrication and measurement Mechanical exfoliation, dry-transfer stacking, electron-beam lithography, etching, SEM/EDX, cryogenic magnetotransport to 12 T, SQUID and NV magnetometry, VNA-FMR to 26 GHz.
Instrument control Python, PyQt6, pyqtgraph, PyVISA, PyMeasure, h5py, MATLAB and LabVIEW where the hardware demands it. Oxford Mercury iPS, ITC 503, ILM 200, Lakeshore 335, Keithley 6221/2182A.
Analysis and modelling NumPy, SciPy, pandas, matplotlib, Jupyter, scikit-learn, Keras/TensorFlow, VASP.
Retrieval and agents PostgreSQL with pgvector, Docling, GROBID, SciNCL embeddings, MCP servers, litellm, Claude Code.
Engineering practice pytest, pytest-qt, mypy, Git, GitHub Actions, Astro for the site you are reading about.
| 2022 – 2026 | PhD candidate, experimental physics, Kläui Group, JGU Mainz. Defence expected by September 2026. |
| 2024 – 2025 | Elected Supply Chain Manager for antiferromagnetic and altermagnetic spintronics in TRR 173 Spin+X, a DFG transregional consortium. Chaired monthly meetings for 20+ researchers across two universities. Co-led the Seedfund proposal that won the 2026 consortium peer vote and is now funded. |
| 2022 – 2025 | Co-organizer and Mainz contact for the Peking-Mainz exchange programme on 2D heterostructures, four bilateral meetings, funded jointly by DFG and NSFC. |
| 2021 – 2022 | Visiting researcher, Kläui Group, JGU Mainz. |
| 2016 – 2021 | Integrated MSc, physics major with computer science minor, NISER Bhubaneswar. DST INSPIRE Scholar fellowship. MSc thesis on spin injection into an organic semiconductor. CS capstone applying neural networks to conformal field theory, ml-cft. |
Outside the lab I have volunteered at Wissenschaftsmarkt Mainz in 2024 and 2025, demonstrating magnetism with open-box experiments, and taught the experimental physics lab courses at JGU across three semesters. I play chess, and I run.
Website · Google Scholar · LinkedIn · sadityakr@gmail.com
If we met at a conference and forgot to exchange contacts, write me a hello.