NABA HASSAN

- Architected enterprise AI systems using LangGraph and optimized RAG pipelines to maintain a 1.18s average response latency on Azure A100 GPU infrastructure. - Resolved a critical 60% system latency bottleneck by re-engineering prompt constraints, execution graphs, and model sampling variables for the Qwen series. - Maintained core R&D retrieval pipelines featuring Reciprocal Rank Fusion (RRF), sparse-dense vector filter parity, and cache validation via BM25IndexCache. - Engineered a high-confidence overrule bypass for documents with Reranker scores above 0.90 to eradicate complex grounding conflicts. - Hardened the security and performance of FastAPI servers by mitigating 11 critical and 9 high-risk vulnerabilities through true Redis-backed OAuth2 token isolation. - Developed a multi-tier RAG Performance Tester suite to validate system constraints under extreme stress conditions. - Ported intent-aware 'Bucket Scaling' token budgeting logic and universal thinking variables to core AI codebases supporting vLLM and llama.cpp backends. - Built automated multi-stage content workflows and onboarding automations using n8n orchestration.

Postgraduate Diploma

NED University of Engineering and Technology

December 2025 - Present

Field: Data Science with Artificial Intelligence

Core Focus: Deep Learning, Neural Networks (ANN/CNN/LSTM), Knowledge Based Systems, Business Intelligence, Ethical Hacking, Digital Forensics, Data Visualization.