const yousra = {
role: "Visual Computing & AI Engineer",
level: "M1 Student — Computer Science (Visual Track)",
location: "🌍 Somewhere between pixels, vectors, and tensors",
currentFocus: [
"Computer Vision & Image Processing",
"Generative AI & Diffusion Models",
"Prompt Engineering & LLM Pipelines",
"3D Rendering & Computational Graphics"
],
philosophy: {
code: "Write it clean, make it beautiful, ship it working",
ai: "Prompts are the new code — craft them with precision",
visuals: "Every pixel is a decision. Make it intentional.",
learning: "Curiosity is the best compiler",
coffee: "Required dependency ☕"
},
dailyRoutine: [
"☕ Boot up with coffee",
"🎨 Explore visual algorithms",
"🧠 Train models, debug intuitions",
"✍️ Craft prompts like poetry",
"📚 Read one paper a day",
"🌙 Dream in latent space"
]
};
🧠 RAG Pipeline — Document Intelligence System
Retrieval-Augmented Generation — Connecting raw documents to meaningful answers
Architecture: Vector Store + LLM Chain
Embedding: Sentence Transformers
Retriever: ChromaDB / FAISS
LLM Backend: OpenAI / Mistral / LLaMA
Status: Active DevelopmentStack:
Highlights:
- 📄 Multi-format document ingestion (PDF, MD, DOCX)
- 🔍 Semantic search with cosine similarity
- 🤖 Context-aware LLM responses
- ⚡ Sub-second retrieval latency
🎨 Visual Prompt Studio — Prompt Engineering Toolkit
Prompt Optimization Platform — Craft, test, and iterate prompts with precision
Purpose: Systematic prompt engineering & evaluation
Interface: React + Tailwind visual UI
Backend: FastAPI + LLM provider abstraction
Feature: A/B prompt comparison & scoringStack:
🖼️ Neural Style Transfer — Visual Computing Project
Real-Time Artistic Style Transfer — Merging content and aesthetics through deep learning
Model: VGG-19 Feature Extraction
Loss: Content Loss + Gram Matrix Style Loss
Interface: Streamlit Web App
Output: High-res stylized imagesStack:
+ Computer Vision & Image Processing ████████████████████░ 92%
+ 3D Rendering & Scene Understanding ████████████████████░ 87%
+ Generative Models & Diffusion ████████████████████░ 85%
+ Computational Photography ████████████████████░ 88%+ Deep Learning & Neural Architectures ████████████████████░ 90%
+ NLP & Large Language Models ████████████████████░ 88%
+ Prompt Engineering & RAG Systems ████████████████████░ 94%
+ Model Fine-Tuning & Evaluation ████████████████████░ 83%+ API Design & Backend Systems ████████████████████░ 85%
+ Frontend & Data Visualization ████████████████████░ 82%
+ DevOps & Containerization ████████████████████░ 78%
+ Research & Technical Writing ████████████████████░ 91%On Visual Computing:
- "The world is a dataset — learn to see it differently"
- Pixels encode meaning. Geometry encodes structure.
- Every model is a lens. Every prompt is a question.
On AI:
- Intelligence is pattern recognition at scale
- Prompts are spells — precision is power
- Align first, optimize second
On Code:
- Write it for humans, optimize it for machines
- Readable today, maintainable forever
- Test early, test often, ship with confidence
