Machine Learning & AI Engineer | GenAI Enthusiast | Cloud Practitioner

I design and build machine learning systems that turn complex data into clear, actionable results. From forecasting stock trends with LSTMs to building multilingual sentiment analyzers and GenAI-powered legal tools, my work focuses on solving real-world problems with clean, scalable AI solutions.

I’ve fine-tuned LLMs, deployed full-stack AI apps, and worked across both classical ML-DL and modern Generative AI pipelines. Currently, I'm exploring the intersection of open-source LLMs and retrieval-augmented generation systems.

☁️ Oracle Certified Generative AI Professional | 🎯 Focused on Enterprise-Ready AI

• Underwater Debris Detection & Analysis

  • Developed a Gradio web application for real-time underwater debris analysis using YOLOv8 computer vision models.
  • Implemented object detection and tracking with ONNX and PyTorch, and a custom ByteTrack-based tracker to accurately count unique trash items in videos.
  • Generated comprehensive outputs including annotated images/videos, statistical summaries, and data visualizations using Pandas, Matplotlib, and Gradio plots.

• ⚖️ GenAI Legal Assistant

  • Fine-tuned FLAN-T5-base on 18,949 legal documents, achieving 40% improvement in legal domain summarization.
  • Integrated BERT-based NER, PyPDF2, and custom tokenization for entity extraction from plain and scanned PDFs.
  • Combined LangChain, OpenAI API, and ChromaDB to build a legal RAG-based assistant.
  • Implemented context-aware document retrieval with FAISS.
  • Deployed app with PDF upload, entity visualization, and download support.
  • Deployed fine-tuned model to Hugging Face Hub; fine-tuned model gained 90+ downloads in 1 month.

• 🌍 Multilingual Sentiment Analyzer

  • Developed a robust ABSA (Aspect-Based Sentiment Analysis) pipeline using PyABSA.
  • Supported multilingual; deployed as a Gradio web app.
  • Processed five benchmark datasets via pandas pipeline, generating structured predictions and confidence scores.
  • Deployed Gradio-based web app with robust input validation and error handling.

• 📃 RAG-Based Academic Policy Assistant

  • Developed RAG chatbot leveraging LangChain and ChromaDB to automate academic policy Q&A with real-time.
  • citation-backed responses reducing student response time from minutes to seconds.
  • Integrated Groq API (LLaMA-3-8B) with custom prompt templates and semantic search for context-aware responses with source attribution and page-level citations.
  • Deployed scalable Streamlit frontend on Hugging Face Spaces with session caching for fast, multi-user access.

• 🟠 Oracle Cloud Infrastructure 2025 Generative AI Professional
• 🧠 5-Day Gen AI Intensive (Kaggle)
• 🤖 Build and Deploy ML Solutions on Vertex AI
• 🧮 Create ML Models with BigQuery ML
• 🚀 Fall in Love with DeepSeek Hackathon

• 📄 Modern Warfare: from boots to bots
• 🧠 Alif 1.0: Pakistan’s First Urdu Large Language Model
• 🤖 Advancements in AGI: A 2025 Perspective

Feel free to reach out if you want to talk AI, collaborate on a project, or just say hi!

• 📨 Email: syedaunrazarizvi3@gmail.com
• 🔗 LinkedIn: linkedin.com/in/aunraza
• 🤗 Hugging Face: huggingface.co/aunraza09