AI Engineer / Data Science graduate focused on LLM applications, RAG systems, and practical AI solutions.
I build modular AI systems that combine retrieval, reasoning, and scalable data pipelines to solve real-world information problems.

• Interested in: GenAI, Agentic RAG, NLP, AI Systems Engineering
• Focused on building portfolio-driven, production-oriented AI projects
• Currently exploring retrieval optimization, metadata-aware search, and evaluation for LLM systems

• Built modular Financial RAG systems for SEC 10-K document QA with citation-backed retrieval
• Experienced with hybrid retrieval, reranking, vector databases, and agentic workflows
• Strong interest in practical AI engineering, evaluation pipelines, and scalable GenAI applications
• Passionate about turning research concepts into deployable AI products

Tech: Python, Qdrant, SentenceTransformers, FastAPI, LLMs, Docker

• Built a modular Financial RAG pipeline for SEC 10-K document QA with citation-backed answers over Apple filings (2021–2025)
• Implemented metadata-aware retrieval, cross-encoder reranking, and agentic retrieval workflows using Query Planner and Evidence Sufficiency Checker
• Improved retrieval performance from Hit@1: 0.54 → 0.88 and MRR: 0.69 → 0.93 across 52 financial QA queries with ~4.9s average latency

Tech: Python, BM25, Qdrant, Streamlit, Knowledge Graphs, ILP Solver

• Built an academic guidance chatbot combining GenAI-based RAG for course lookup and algorithmic reasoning for curriculum planning
• Developed a Hybrid RAG pipeline using BM25 + Qdrant, improving course lookup accuracy from 29% → 76%
• Designed an algorithmic reasoning engine using Knowledge Graphs and ILP solvers to generate valid multi-year study plans with 100% prerequisite constraint accuracy

Python SQL

NLP OCR LLMs RAG Machine Learning Deep Learning

PyTorch TensorFlow/Keras scikit-learn SentenceTransformers

Qdrant PostgreSQL Docker Apache Kafka

Git FastAPI Flask Streamlit REST APIs

• Agentic RAG architectures for financial document intelligence
• Retrieval evaluation and metadata-aware search optimization
• LLM orchestration and low-latency AI pipelines
• Production-ready GenAI applications with scalable backend systems

• LinkedIn: https://www.linkedin.com/in/nxvinh1907/
• GitHub: https://github.com/vinhnx0
• Email: [email protected]

Building practical AI systems that combine retrieval, reasoning, and scalable data engineering for real-world applications.