Research focus
LoRA, QLoRA, code generation, debugging, code explanation, and efficient inference.

Exploring efficient adaptation of suitably licensed open-weight models for code understanding and generation.
LoRA, QLoRA, code generation, debugging, code explanation, and efficient inference.
Establish data licensing, baseline benchmarks, reproducible fine-tuning, and safety evaluation before any release.
Model selection → baseline evaluation → adapter training → benchmark reporting.