Research & Applications
Research topics, framed as transferable capabilities.
I do not want this blog to be boxed into one thesis topic. These categories keep the research side broad: multilingual systems, evaluation, responsible AI, retrieval, and applied ML engineering.
Research & application tracks
These labels are designed to cover current work and leave room for future AI/NLP directions.
Multilingual AI
Machine translation, multilingual NLP, low-resource adaptation, LoRA domain NMT, domain terminology, language variation, and cross-lingual evaluation.
Model Evaluation
Metric design, benchmark construction, paired comparisons, statistical testing, ablations, error analysis, and human evaluation.
Explainability & Responsible AI
Bias evaluation, interpretability, training-data attribution, Shapley analysis, controlled prompts, model auditing, and responsible data practices.
Retrieval & Knowledge Systems
RAG, search, sparse/dense/hybrid retrieval, graph retrieval, long documents, evidence grounding, and document QA.
Applied ML Systems
Data analysis, classification, forecasting, information extraction, ML pipelines, deployment, demos, and practical AI products.