.png&w=256&q=75)
Shehzad Ahmed@Shehzad786
3
Events attended
1
Submissions made
Pakistan
7+ years of experience
About me
Shehzad Ahmed is an AI/ML Engineer with strong expertise in machine learning, deep learning, and computer vision. He specializes in building intelligent systems using Python and modern AI frameworks. His technical skill set includes machine learning algorithms, data preprocessing, feature engineering, deep learning architectures, and software development practices. Shehzad has worked on multiple AI-driven projects including document automation systems using OCR, fine-grained bird classification using deep learning models, and object detection systems using YOLO models. He has experience working with advanced architectures such as DenseNet and convolutional neural networks for image classification tasks. In addition to computer vision, he has experience with Natural Language Processing (NLP) and Retrieval-Augmented Generation (RAG) systems. He has worked with vector databases like Weaviate, open-source models from Hugging Face, and frameworks such as LangChain to build intelligent question-answering systems over code repositories. Shehzad is also interested in open-source large language models (LLMs) and their integration into real-world applications. His work focuses on developing scalable AI solutions that automate workflows, improve data processing, and deliver intelligent insights. He continuously explores new technologies in artificial intelligence, computer vision, and NLP to build innovative solutions and contribute to modern AI development.
🤓 Latest Submissions
LexiAgent — AI Legal Expert
LexiAgent is an autonomous multi-step AI agent that solves a real enterprise problem: most businesses sign legal contracts without fully understanding the risks. Lawyers charge $200–500/hour. LexiAgent does the same job in seconds, for free. Upload any contract in PDF, DOCX, XLSX, PPTX, HTML, or TXT format. The LangGraph-powered pipeline automatically kicks in — no human intervention required at any step. The agent runs through 5 autonomous nodes: First, the Document Loader reads the file using Docling, handling any format. Second, the Contract Parser uses Claude to extract all clauses, identify parties, and pull key dates. Third, the Risk Analyzer scores every clause from 0 to 100 and flags HIGH, MEDIUM, and LOW risks — with a conditional edge that loops back for deeper analysis if the overall risk score is critically high. Fourth, the Negotiation Advisor generates professional rewrite suggestions for every risky clause, complete with negotiation tips. Fifth, the Report Generator compiles a full executive summary that can be downloaded as PDF or TXT. After analysis, users can chat directly with their contract through the AI Advisor tab — asking questions like "Should I sign this?", "What is the riskiest clause?", or "How do I negotiate clause 3?" The AI answers with full context of the analyzed contract. The codebase follows SOLID principles throughout — Single Responsibility per node, Open/Closed BaseNode for extensibility, Dependency Inversion via LLMClient abstraction. Built with FastAPI async backend and Server-Sent Events for real-time streaming pipeline updates to the frontend.
19 May 2026
