Saad Shakeel
Transforming ideas into intelligent AI solutions that connect technology with human experience, building the future of interaction through Generative AI and intelligent automation.
Work Experience
Generative AI Developer (R&D)
Advanced Telecom Services (ATS) AI Lab
- Conducted R&D in Generative AI and emerging LLM technologies, exploring new trends, frameworks, and practical use cases to identify innovative approaches to AI development.
- Experimented with RAG, AI agents, and LLM-powered applications to build prototypes, evaluate emerging techniques, and explore practical solutions for real-world AI use cases.
AI Developer
Robotics World Pvt. Ltd
- Focused on developing AI-powered applications using LLMs, Retrieval-Augmented Generation (RAG), and AI agents, handling the complete development lifecycle from system design to deployment.
- Built and deployed production-ready AI solutions using FastAPI, with applications hosted on Hostinger and Google Cloud Platform (GCP).
Data Science Intern
Advanced Telecom Services (ATS) AI Lab
- Gained practical experience with Microsoft Azure, Big Data, and advanced data analysis techniques.
- Deployed scalable, cloud-based solutions and explored applications of Large Language Models (LLMs) in real-world scenarios.
Education
Bachelor of Science in Computer Science
Abbottabad University Of Science & Technology
FSC Pre-Engineering
Govt Postgraduate College No 1 Abbottabad
Certifications & Skills
Certifications
Skills
Featured Projects
Steam-Minds (AI Modules)
AI-powered EdTech platform featuring course-specific AI tutors, RAG-based document search, automated transcript verification, and AI quiz generation.
Steam-Minds (AI Modules)
AI-powered EdTech platform featuring course-specific AI tutors, RAG-based document search, automated transcript verification, and AI quiz generation.
Key Architecture Highlights
- Built course-specific AI tutors that understand course content and provide relevant answers using conversational memory.
- Architected high-accuracy document and syllabus retrieval powered by Qdrant vector database
- Improved existing AI modules for automated transcript verification and curriculum-based quiz generation.
- Developed and deployed FastAPI-based AI services on a Linux VPS for production use.
Uplaw (Legal AI Assistant)
Legal AI ntelligence assistant built for the Uplaw legal platform, indexing 2,500+ legal statutes and case documents using hierarchical Parent-Child chunking, hybrid vector retrieval on Qdrant Cloud, and automated CI/CD deployment on Google Cloud Run.
Uplaw (Legal AI Assistant)
Legal AI ntelligence assistant built for the Uplaw legal platform, indexing 2,500+ legal statutes and case documents using hierarchical Parent-Child chunking, hybrid vector retrieval on Qdrant Cloud, and automated CI/CD deployment on Google Cloud Run.
Key Architecture Highlights
- Ingested and structured 2,500+ statutory legal documents using advanced Parent-Child chunking for maximum contextual fidelity
- Combined dense vector embeddings with sparse keyword search on Qdrant Cloud for ultra-accurate legal precedent retrieval
- Containerized FastAPI microservices deployed on Google Cloud Run with automated GitHub Actions CI/CD pipelines
Stock Market Agent
AI-powered financial agent and interactive dashboard for real-time and historical S&P 500 market analysis using LangGraph and Yahoo Finance APIs.
Stock Market Agent
AI-powered financial agent and interactive dashboard for real-time and historical S&P 500 market analysis using LangGraph and Yahoo Finance APIs.
Key Architecture Highlights
- Autonomous agent reasoning over financial datasets and stock tickers
- Real-time technical indicator calculation and interactive chart visualization
- Conversational market Q&A interface powered by LLM agent tool use
AI Trip Planner
Autonomous AI travel assistant that uses real-time Google Places and Maps APIs through LangGraph agent workflows to generate optimized itineraries.
AI Trip Planner
Autonomous AI travel assistant that uses real-time Google Places and Maps APIs through LangGraph agent workflows to generate optimized itineraries.
Key Architecture Highlights
- Multi-step agent planning for real-time location, hotel, and attraction discovery
- Interactive conversational UI built with Streamlit and Groq LLM API
- Dynamic route mapping, weather-aware suggestions, and itinerary export
Agentic-RAG Notebooks
A collection of advanced RAG implementations with agent-based workflows, featuring vector search, hybrid retrieval, and self-corrective retrieval.
Agentic-RAG Notebooks
A collection of advanced RAG implementations with agent-based workflows, featuring vector search, hybrid retrieval, and self-corrective retrieval.
Key Architecture Highlights
- Implementation of Self-RAG and Adaptive RAG agent patterns
- Hybrid retrieval combining dense vector search with sparse BM25 indexing
- Comparative benchmarking across cloud vector stores and local embeddings
Emotion-Driven Music App
Real-time facial emotion recognition pipeline that analyzes webcam video streams to deliver personalized, mood-based music recommendations.
Emotion-Driven Music App
Real-time facial emotion recognition pipeline that analyzes webcam video streams to deliver personalized, mood-based music recommendations.
Key Architecture Highlights
- Real-time facial landmark and emotion detection via OpenCV and CNN models
- Asynchronous FastAPI backend for low-latency recommendation inference
- Dynamic playlist generation matched to detected user mood states
PakCurrency-VisionAI
Deep learning computer vision system for Pakistani currency denomination detection and counterfeit verification using custom YOLO object detection models.
PakCurrency-VisionAI
Deep learning computer vision system for Pakistani currency denomination detection and counterfeit verification using custom YOLO object detection models.
Key Architecture Highlights
- Custom YOLO model trained on Pakistani banknote dataset
- Multi-class denomination classification and security feature verification
- High-accuracy real-time inference pipeline for mobile and desktop
Zaran Tours Website
Production responsive travel and destination booking platform with modern UI components, fast load times, and fluid mobile navigation.
Zaran Tours Website
Production responsive travel and destination booking platform with modern UI components, fast load times, and fluid mobile navigation.
Key Architecture Highlights
- Mobile-first responsive layout with fast render metrics
- Interactive destination showcase and tour itinerary browser
- Optimized media delivery and accessibility-first markup
Gallery
About Me
I’m focused on building intelligent AI applications that turn ideas into practical solutions and meaningful user experiences.
My work centers around Artificial Intelligence and Generative AI, with a focus on building AI-powered applications, intelligent assistants, automated workflows, and solutions that solve real-world problems.
I enjoy going beyond theory by learning through practical projects, experimenting with new technologies, and exploring how AI can be used to make everyday work smarter and more efficient.
Let’s connect and build something intelligent together.
Let’s Build Something Intelligent Together
I’m always open to discussing AI systems, agentic workflows, RAG architectures, and innovative collaboration opportunities.