h
hifza_ai

Hifza

@hifza_ai

AI Engineer, Generative AI and RAG Systems Specialist

Pakistan
Inglese
Alcune informazioni sono riportate in lingua inglese.
Chi sono
Hi, I'm an AI Developer specializing in RAG (Retrieval-Augmented Generation) systems, LLM applications, and intelligent document Q&A. I turn messy PDFs and unstructured data into clean, searchable knowledge bases using Python, vector search, and LLM integration. Background: 1 year of hands-on Python, ML/DL, and data analysis experience, now focused on RAG and LLM-powered systems. Core Services: - RAG Systems & Document Q&A - LLM Integration (Gemini, OpenAI) - PDF Parsing & Data Extraction - Vector Search (FAISS) & Streamlit Apps Clean code, clear communication. Let's discuss your project!... Continua a leggere

Competenze

h
hifza_ai
Hifza
offline • 

Consulta i miei servizi

Automazioni
I will extract data and tables from PDF files using python
A partire da5 USD / progettoMaggiori dettagli

Portfolio

Esperienza lavorativa

GitHub

AI Engineer (Independent Project) — RAG & Data Extraction Systems

GitHub • Lavoratore autonomo

May 2026 - Present • 5 mos

Built an end-to-end enterprise Retrieval-Augmented Generation (RAG) system (TechCorp Financial RAG) that processes large-scale financial documents and structured data. Key highlights: - Designed a PDF data-extraction pipeline that parsed 250 SEC 10-K financial filings (28,862 pages) into 131,209 clean, source-traceable text chunks - Built a semantic search layer using vector embeddings (FAISS, ChromaDB) to enable accurate retrieval from unstructured documents - Developed a hybrid query engine that routes natural-language questions between a structured SQL database (1M+ rows) and unstructured document search - Implemented a natural-language-to-SQL agent with schema-aware prompting and security guardrails against destructive queries - Optimized database performance using B-Tree indexing, reducing query time by over 99% on large datasets - Deployed a Streamlit web dashboard with a live chat interface and source-traceability panel Tech stack: Python, LangChain, PyMuPDF, FAISS, ChromaDB, SQLite, Streamlit, FastAPI, Google Gemini 2.5 Flash, HuggingFace Embeddings Self-taught AI/ML engineer (BS IT, 8th semester) with hands-on, production-style project experience in data parsing, automation, and AI-driven retrieval systems.