I'm an AI master's student at Boston University building applied ML, data, and backend systems.
I like projects where the hard part is making messy inputs useful: retrieving the right evidence, cleaning real-world data, evaluating models honestly, and turning the result into something people can use.
- Local RAG and grounded scientific QA
- Data science pipelines for messy transaction datasets
- Computer vision and multimodal model analysis
- Backend APIs with Django REST Framework and FastAPI
Scientific QA with Local RAG
Built a local QASPER pipeline with hybrid retrieval, reranking, Phi-3.5 generation, and citation evaluation.
AML Transaction Risk Detection
Engineered 50+ historical behavior features across IBM + SAML-D transaction data for alert ranking.
Video Moment Retrieval Analysis
Reproduced DETR-style video moment retrieval models and analyzed temporal bias, language sensitivity, and attention heads.
DRFChatApp
Django REST Framework chat backend.
Python SQL PyTorch TensorFlow scikit-learn pandas NumPy
Django REST Framework FastAPI JavaScript Git Linux Jupyter
- LinkedIn: linkedin.com/in/zukhriddin-fakhriddinov
- Email: zfmsai@bu.edu

