Leading research on calibration of LLM-based code-completion models inside JetBrains IDEs. Designed user studies measuring how developers accept, trust, and act on AI model suggestions. Collaborated across institutional lines to deliver empirical evidence on AI assistance in real workflows. Co-authored publications and supported data-engineering pipelines for experiment logging, analysis, and visualisation.

Investigate novel architectures and fine-tuning methods for code-specialised language models. Work on integration strategies for AI-assisted development environments—focusing on developer experience, cognitive load, and action sequencing. Collaborate across lab and industry to build evaluation suites and prototype tooling.

Contributed to research on LLMs4Code: built pipelines for data processing, model training, evaluation, and integration. Supported published work on developer–AI interaction and empirical studies. Worked closely with research leads on dataset curation and modeling decisions.

Evaluated synthetic-data generation methods and designed benchmarking experiments. Developed a novel time-series synthetic-data generation method conditioned on metadata, contributing to the product roadmap. Integrated ML research literature into production-oriented pipelines.

Investigated bias-mitigation methods in synthetic-data generation for ML fairness. Applied synthetic-data techniques to design controlled experiments for fairness and reliability in AI. Gained experience in research project planning, result communication, and collaborative development.

Courses: Software Engineering; Algorithms & Data Structures; Machine Learning; Computer Networks; Computer Architecture; Databases; Web Technologies; Data Processing. Expected graduation: Feb 2026.

Topics: Number & Set Theory, Statistics, Differential Equations, Information Theory, Markov Chains.