CS student at Arizona State University building AI/ML tools that make complex technology accessible. I thrive in fast-paced environments—hackathons are where I do my best work—and I'm passionate about creating solutions that enhance the human experience.

• 🔭 Currently working on:
  • CloudForge - Knowledge-graph-driven AWS infrastructure builder with multi-agent LangGraph pipeline
  • VisionForge - Visual neural network builder with drag-and-drop interface and code export
  • LearnLens - Educational webcam filter app teaching JavaScript through 8 progressive levels
  • CodeStream - JupyterLab extension for real-time teacher-student code collaboration
• 🌱 Learning more about distributed systems, multi-agent AI, cloud infrastructure, and computer vision
• 🏆 1st Place Winner at HackASU (Claude Builder Club) for VisionForge - Developer Tools track
• 🏆 2nd Place Winner at DevHacks for Decoy - AI-powered scam call detection system
• 💡 I love discovering new tools and building things that solve real problems
• 🎯 Looking for SWE/AI-ML internships where I can build impactful products

Knowledge-Graph-Driven AWS Infrastructure Builder

• Designed a 5-agent LangGraph pipeline: PRD refinement → architecture planning → Terraform generation → validation → deployment
• Architecture decisions are derived from a knowledge graph of validated cloud patterns via RAG—never free-form LLM output
• Generates production-grade Terraform HCL (main.tf, variables.tf, outputs.tf) with a deterministic validation loop: terraform fmt → validate → TFLint → Checkov
• Zero persistent cloud credentials: uses IAM role assumption (AWS STS AssumeRole) and Fernet symmetric encryption at rest
• Streams every agent stage live via Server-Sent Events (PRD refinement, architecture reasoning, file-by-file generation, deploy logs)
• Commits generated scaffolds directly to the user's GitHub repo via GitHub App OAuth
• Human-in-the-loop interrupts at architecture approval and PRD acceptance stages using LangGraph interrupt() API
• Contributions: Full-stack development, multi-agent architecture, knowledge graph design, security model

Tech: Next.js 15 TypeScript FastAPI Python LangGraph MongoDB AWS Terraform Claude Sonnet Ollama Zustand Tailwind CSS

AI-Powered Precision Farming Platform — Intelligent Planet Hackathon 2025

• Built real-time sensor monitoring dashboard for soil, temperature, humidity, pH, NPK, light, and rainfall data
• Integrated Google Gemini AI for on-demand agricultural advice and field-specific insights
• Implemented interactive field mapping with Google Maps JavaScript API
• Added bilingual (English + Arabic RTL) support and full dark mode with system preference detection
• Designed analytics dashboard with ROI tracking, water usage comparison, and yield progression charts
• Contributions: Full-stack development, AI integration, and UI/UX design

Tech: Next.js 15 TypeScript Tailwind CSS Google Gemini API Firebase Auth Google Maps API Zustand Kendo UI Framer Motion

Chrome Extension + AI Backend to Fight LinkedIn Engagement Farming

• Built a Chrome Extension (Manifest V3) that detects "comment for link" engagement-bait posts on LinkedIn
• Engineered multi-source search pipeline: scrapes author comments, checks profile featured links, falls back to web search
• Used Google Gemini to understand post context and rank link relevance by semantic similarity
• Built caching layer with Redis to avoid redundant searches and enforce rate limits
• Containerized FastAPI backend deployed via Docker; extension is Chrome/Firefox compatible
• Contributions: Solo project—extension UI, backend API, AI pipeline, and scraping logic

Tech: Python FastAPI Google Gemini API Playwright Redis PostgreSQL Chrome Extension (MV3) Docker

Visual Neural Network Builder with Code Export

• Built drag-and-drop interface for designing deep learning architectures without writing code
• Implemented automatic shape inference eliminating manual tensor dimension tracking
• Developed multi-framework export system generating clean PyTorch and TensorFlow code
• Integrated Google Gemini AI assistant for natural language model modifications
• Created real-time validation system catching architecture errors before export
• Supports 20+ layer types including CNNs, LSTMs, ResNets, and custom operations
• Won 1st place in Developer Tools track at HackASU hosted by Claude Builder Club
• Contributions:Full-stack development, AI integration, and UI/UX design

Tech: React TypeScript Django Python Gemini AI Tailwind CSS ReactFlow Vite

AI-Powered Scam Call Detection & Routing System

• Built complete call routing infrastructure in 24-hour hackathon with 4-person team
• Engineered Twilio & Vapi integration that automatically routes suspected scam calls to AI personas
• Developed 4 distinct AI personalities that keep scammers engaged while protecting users
• Created analytics dashboard tracking call duration, persona effectiveness, and system health
• Integrated trained ML models with call routing logic for real-time risk profiling

Tech: Python FastAPI Twilio Vapi SQLite 11labs Machine Learning

Educational Webcam Filter App - Learn JavaScript Through Building

• Created comprehensive JavaScript learning platform with 8 progressive levels
• Designed educational pathway covering React, Canvas API, WebRTC, and Machine Learning
• Built 20+ CSS filters with real-time webcam processing and intensity controls
• Integrated TensorFlow.js for face detection with 68-point facial landmarks
• Implemented Google Gemini Vision API for AI-powered skin analysis and recommendations
• Developed draggable sticker system with coordinate transformations
• Structured project with git branches for each learning level and solution references
• Contributions: Solo project - curriculum design, full implementation, and documentation

Tech: React JavaScript WebRTC Canvas API TensorFlow.js Gemini Vision API Face-API.js

JupyterLab Extension for Real-Time Collaborative Learning

• Built complete Redis-powered backend infrastructure for selective cell synchronization
• Developed responsive frontend interface with cell-level sync controls and preview overlays
• Enables teachers to push code updates while maintaining student autonomy over when to sync
• Solving real classroom pain points I've experienced as both student and workshop mentor
• Contributions: Frontend development, landing page design, and WebSocket integration

Tech: Python JavaScript Redis WebSockets JupyterLab Docker Tornado

No-Code LLM Fine-Tuning Platform

• Developed GPU-optimized platform enabling non-technical users to fine-tune LLMs locally
• Implemented hardware-aware algorithms auto-configuring training parameters based on GPU/VRAM
• Built responsive React frontend simplifying complex ML concepts (PEFT, LoRA) for accessibility
• Supports text generation, summarization, and extractive Q&A tasks
• Delivered complete prototype in 48 hours at DevHacksXStrategy hackathon
• Contributions: Complete frontend UI/UX design and implementation

Tech: Python React FastAPI PyTorch Transformers LoRA bitsandbytes

Network Security Monitor with ML Anomaly Detection

• Designed and developed frontend interface for real-time network packet capture and analysis
• Created intuitive visualizations for TCP, UDP, ICMP traffic and protocol-specific anomaly metrics
• Led frontend development in 12-hour hackathon with seamless backend integration
• Contributions: Frontend development and real-time data visualization

Tech: Python React TensorFlow Scapy Network Security

Contributing to multiple open-source projects focused on education technology and AI tools:

• CodeStream - JupyterLab notebook synchronization extension
• ModelForge - No-code LLM fine-tuning platform
• NetworkAnalyser - ML-powered network security monitoring

Participated in hackathon building solutions for non-profits and community organizations:

• NMTSA Education Platform - Contributing to educational technology solutions

• ☁️ Building CloudForge - knowledge-graph-driven AWS infrastructure builder with a 5-agent LangGraph pipeline
• 🎨 Improving VisionForge - visual neural network builder revolutionizing deep learning prototyping
• 🔨 Enhancing CodeStream to revolutionize classroom coding workshops
• 🤖 Exploring multi-agent AI architectures, RAG pipelines, and distributed systems
• 📚 Contributing to open-source projects in AI/ML and education technology
• 🎯 Participating in hackathons and building innovative solutions
• 💡 Always on the lookout for the next cool tool or framework to experiment with!

• ♟️ Playing chess (formerly Tech Manager at PeelChess nonprofit)
• 🎮 Gaming and exploring new indie titles
• 📚 Reading sci-fi and tech philosophy
• 🚴‍♂️ Cycling around campus
• ⚽ Playing football, basketball, and hockey
• 🎵 Discovering new music

I'm always excited to collaborate on projects that make technology more accessible and solve real-world problems. Whether it's AI/ML, full-stack development, or just a cool idea you want to build—let's chat!