m.s. cs @ texas tech university.
interested in nlp, rag, deep learning, and building end-to-end ml systems on real data.
currently thinking about retrieval quality in rag, code-mixed language models, and cloud-based ml workflows.
legal document intelligence platform
multi-user rag platform for legal document q&a with cited answers.
hybrid retrieval (vector search + bm25 with reciprocal rank fusion), ocr fallback, jwt auth, async processing.
fastapi chromadb openai celery redis
code
cyberbullying detection in telugu
built a 37.8k-sample telugu–tenglish dataset and benchmarked seven bert-based models under identical conditions.
95.3% accuracy with mbert; distilbert was the most efficient for deployment.
pytorch transformers indicnlp scikit-learn
code
skin cancer detection with cnns
deep learning classifier for skin cancer images using resnet and fastai.
includes the dataset, notebook, pr and roc curves, confusion matrix, and a literature review.
fastai resnet jupyter
code
- rag systems and retrieval quality
- multilingual and code-mixed text (telugu, english)
- evaluation and error analysis for nlp models
- building end-to-end systems on real data
languages: python · java · javascript · c · sql
ml / dl: pytorch · tensorflow/keras · hugging face · fastai · scikit-learn · numpy · pandas
backend & web: fastapi · rest apis · celery · redis · react (basics) · streamlit
cloud & tools: aws (fundamentals) · git · linux
- aws academy cloud architecting · view
- aws academy graduate – cloud foundations · view
- oci generative ai professional · view
- oci ai foundations associate · view
- cs50's introduction to programming with python · view
- mern full stack development program · view
piano, video games, sports and fitness.