Bachelor of Science in Computer Science
Government College University, Lahore
Focused on core computer science foundations, problem solving, and software development.
Available for AI/ML roles
Full Stack AI Engineer
Javaid Butt
Building practical AI products with RAG, LLM integrations, and clean web interfaces.
Bridging the gap between LLM research and production-ready web applications
1+ Year Experience
15+ Projects Delivered
4+ LLM Platforms
About
I build production-grade AI products that combine robust retrieval systems with intuitive user experiences. My work spans end-to-end architecture from data ingestion and embedding pipelines to deployment-ready web interfaces.
I specialize in RAG pipelines, practical LLM integration across DeepSeek, Mistral, Ollama, and Perplexity, and scalable full-stack delivery using Next.js and FastAPI.
Over the last 1 year, I have delivered AI-driven platforms for fintech, operations, and enterprise knowledge workflows where reliability, observability, and latency matter.
1+ Year Experience
15+ Projects Delivered
4+ LLM Platforms
Education
Academic and leadership credentials that support my technical work in AI and software projects.
Aspire Leadership Programme
Harvard Business School
Leadership, communication, and decision-making training for high-impact professional growth.
Skills
AI / ML
PyTorch
LangChain
LlamaIndex
OpenAI API
RAG
Vector DBs
HuggingFace
Backend
Python
FastAPI
Node.js
PostgreSQL
Redis
Docker
REST APIs
Data Science
Pandas
NumPy
Matplotlib
Seaborn
Scikit-learn
Streamlit
What I Offer
Services
RAG System Development
Design and implementation of retrieval-augmented generation pipelines with document chunking, embeddings, and citation-aware responses.
- • Semantic retrieval
- • Vector database setup
- • Answer grounding
LLM Integration
Integration of DeepSeek, Mistral, Ollama, and Perplexity APIs into practical workflows for chat, summarization, and automation.
- • Prompt workflows
- • Model routing
- • API orchestration
Data Science Solutions
End-to-end data analysis and ML prototypes using pandas, NumPy, scikit-learn, and Streamlit for actionable insights.
- • EDA and preprocessing
- • Model experiments
- • Interactive dashboards
AI Workflow Automation
Automate repetitive business and analysis tasks with robust Python pipelines and lightweight service integrations.
- • Pipeline scripting
- • Task automation
- • Deployment-ready flows
Projects
Tour Guide Web Portal
A complete travel planning and destination discovery portal with a clean user experience for exploring guides, routes, and key trip information.
Next.jsTypeScriptTailwind CSSVercel
Credit Card Fraud Detection
Binary classification project to detect fraudulent vs non-fraudulent credit card transactions using a Kaggle dataset with focused preprocessing, class imbalance handling, and model evaluation.
PythonPandasScikit-learnKaggle Dataset
AI Avatar Call Agent
Full-stack AI-powered voice and video calling platform with realistic avatar rendering and intelligent support automation. Users converse with an AI agent in real time while an avatar streams on screen.
OpenAI Realtime APILiveKitNext.jsFastAPIBeyond Presence
Robotics Vision: YOLO-World & YOLOE Fine-Tuning
Fine-tuned YOLO-World and YOLOE on a custom robotics dataset (hands, gloves, metal parts) across multiple video sources using domain-specific annotations on Google Colab T4 GPU. Ran comparative inference pipelines producing annotated video outputs with evaluated detection precision.
YOLO-WorldYOLOEPythonGoogle ColabComputer Vision
Open Record
Platform that fetches documents from the US Federal Register, summarizes each document, surfaces related similar orders, and includes a chatbot for interactive question answering on the retrieved content.
PythonFastAPIRAGLLM Chatbot
Potato Disease Classification
End-to-end deep learning system that classifies potato leaf diseases (Early Blight, Late Blight, Healthy) with a FastAPI backend, React web interface, React Native mobile app, and GCP deployment.
TensorFlowFastAPIReactReact NativeGCP