Research Fellow — Computational Materials + ML Systems at University College London, Department of Chemistry (2025-10 – Present)
Supervisor: Prof. Scott Woodley.
- Co-developer of active-learning and surrogate/emulator workflows for accelerated materials discovery in collaboration with the Alan Turing Institute through AutoEmulate, linking uncertainty aware acquisition with rapid model iteration.
- Built end to end scientific pipelines connecting atomistic simulation outputs to ML-ready datasets, model evaluation loops, and scalable high throughput exploration of inorganic materials, surfaces, interfaces, and nanoclusters.
- Designed and maintain UnifiedLab, an open source Rust meta scheduler for software agnostic mixed GPU/CPU workflows across heterogeneous HPC environments, improving reproducibility, portability, and operational efficiency of computational materials pipelines.
- Built workflow infrastructure integrating hybrid DFT and atomistic simulation codes, enabling robust orchestration of high throughput calculations across local, cloud, and HPC environments.
- Prototype AI-assisted workflow components for scientific decision support, including LLM integration and Gemma fine tuning on DGX Spark for rapid experimentation in AI for materials.
- Collaborate with chemists, physicists, and software-focused partners across the Materials Chemistry community to translate research requirements into deployable computational tooling.
Research Associate — Computational Materials Discovery at Imperial College London, Department of Chemistry (2022-10 – 2025-09)
Supervisor: Prof. Kim E. Jelfs. Research contributed to seven high-impact publications.
- Built generative search pipelines using evolutionary algorithms for molecular and materials candidates under property driven objectives, reducing search cost by an order of magnitude relative to brute force exploration.
- Designed high throughput discovery workflows connecting candidate generation, simulation based property evaluation, ranking, and iterative refinement for porous organic cages and host guest systems.
- Enabled synergistic computational/experimental discovery of high symmetry porous organic cages from low symmetry precursors through multi scale computational screening.
- Developed, validated, and maintained research software for materials discovery using quantum chemistry and atomistic simulation methods.
- Worked across theory and experiment to translate computational results into actionable molecular design targets and discovery campaigns.
Graduate Research Assistant (PhD Researcher) at Koç University, Materials Science & Engineering (2016-09 – 2021-12)
Thesis: High-Throughput Computational Screening of MOFs for Carbon Dioxide Capture and Hydrogen Purification. Supervisor: Prof. Seda Keskin Avcı. Research contributed to eight high impact publications.
- Performed atomistically detailed simulations on 10,000+ MOFs, identifying promising next-generation adsorbents and membranes for CO2 capture and H2 purification.
- Conducted structure property analysis across 50,000+ Grand Canonical Monte Carlo and Molecular Dynamics simulations.
- Built large-scale datasets and screening workflows for crystalline framework materials, supporting benchmarking and ML ready structure-property analysis.
- Investigated the influence of metal site characteristics on gas sorption and diffusion in MOFs, helping inform new experimental strategies.
Graduate Research Assistant (MSc Researcher) at Sabancı University, Materials Science & Engineering (2013-09 – 2015-08)
Thesis: Investigation of machinability of green bodies of Solid Cast Alumina through the Addition of a Poly (Carboxylate Ether)-based Superplasticizer. Supervisors: Asst. Prof. Özge Akbulut; Prof. Yusuf Z. Menceloğlu.
- Formulated machinable ceramic green bodies via addition of PCE-based rheology modifiers; outcomes contributed to a European patent and a first-author paper.
