// AI Engineer

Mohamed Abdalkader

> AI Engineer

LLM Fine-tuning · Computer Vision · Medical AI · MLOps

AI Engineer with 1+ year building production ML systems for medical imaging and NLP. Delivered 240% accuracy gains via LoRA fine-tuning and multi-agent architectures. Deployed RAG pipelines indexing medical textbooks with sub-second retrieval. Reduced model memory 59% through quantization.

View Projects → Read Blog ↓ Download CV
Mo-Abdalkader ~ $ 
class MohamedAbdalkader:
  role   = "AI Engineer"
  focus  = [
    "LLM Fine-tuning",
    "Computer Vision",
    "Medical AI",
    "RAG Systems",
  ]
  results = {
    "accuracy_gain"  : "+240%",
    "memory_savings" : "-59%",
    "grants"         : "$3,450",
    "caption_boost"  : "26x",
  }
  status = "open_to_opportunities"
📈 240% Accuracy Gain
⚡ 59% Memory Reduction
🏆 $3.4K Research Grants
🚀 8+ ML Projects Shipped
// featured work
Live LLMs

NeuralChat — Multi-Provider LLM Chatbot

Multi-provider LLM chatbot supporting 10+ models (Cohere, OpenAI, Groq) with real-time cost & latency display. 5 prompting strategies, 6 AI personas. Deployed on Railway.

FastAPILangChainGroqRailway
⌥ GitHub ↗ Live Demo
Live LLMs

Mo-Bot — Personal AI Telegram Bot

Serverless AI assistant on Cloudflare Workers. Three-model fallback chain (LLaMA 3.3 70B → 3.1 70B → 3.1 8B) across Groq. Bilingual Arabic/English with Arabizi detection.

Cloudflare WorkersGroqLLaMA 3.3Telegram API
↗ Live Demo
Private Research Medical AI

Medical VLM with Multi-Agent Architecture

Fine-tuned Qwen 2.5 7B VLM with LoRA rank-32 on 35,000 retinal images. Multi-agent pipeline achieved 240% accuracy boost (20% → 68%). 4-bit QLoRA cut VRAM 17GB → 7GB.

Qwen 2.5 7BPyTorchLoRA/QLoRAFAISS
🔒 Private
Private Research Medical AI

Medical Knowledge RAG System

RAG pipeline indexing 20+ medical textbooks (5,000+ pages) with FAISS. 0.3s retrieval latency, 0.89 recall@5. Grounded VLM outputs lifted retinopathy to 68%, edema to 94%.

LangChainFAISSTransformerssentence-transformers
🔒 Private
Live Computer Vision

Brain Tumor Classification

Custom CNN — 95% accuracy on 3,000 brain MRI scans across 4 tumor types. 400% dataset expansion via augmentation. Streamlit app with Class Activation Maps.

TensorFlowKerasOpenCVStreamlit
⌥ GitHub ↗ Live Demo
Live Computer Vision

Face Recognition & Similarity System

FaceNet-based recognition with SQLite identity database managing 1,000+ profiles. 0.92 accuracy at 15 FPS on CPU. Desktop Tkinter GUI plus Streamlit web API with cosine similarity matching.

PyTorchFaceNetStreamlitSQLiteOpenCV
⌥ GitHub ↗ Live Demo
Live Research

AI-IoT Renewable Energy Prediction

Graduation project benchmarking 21 DL architectures (LSTM, GRU, BiLSTM, CNN-LSTM, Transformer, TFT) for solar and wind forecasting. 98% accuracy, <15% MAPE, 48-hour window. ITAC + ASRT funded. IEEE national recognition award.

LSTM · GRUTransformersAzure IoT HubTensorFlow
⌥ GitHub
View All Projects →
// latest posts

Blog

Deep dives, tutorials, and thoughts on AI.

Day 1 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 19, 2026 · 11 min read

What is Machine Learning?

Machine learning explained from zero — what it is, how it differs from normal programming, the three types, and why it now powers everything.

Read more →
Day 2 🌙 30-Day AI Ramadan Series
Practical Guide · Feb 20, 2026 · 10 min read

The ML Pipeline

The complete journey from raw data to a deployed model — data collection, cleaning, feature engineering, training, evaluation, and deployment.

Read more →
Day 3 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 21, 2026 · 12 min read

How Models Actually Learn

Loss functions, gradient descent, and learning rate — the universal engine behind every ML model from linear regression to GPT-4.

Read more →
Day 4 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 22, 2026 · 10 min read

Linear Regression

Linear regression from scratch — the simplest and most foundational ML model. OLS, MSE, R², and why it still powers production systems.

Read more →
Day 5 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 23, 2026 · 10 min read

Logistic Regression

How a linear model learns to output probabilities using the sigmoid function, log-loss, and what the decision boundary really means.

Read more →
Day 6 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 24, 2026 · 10 min read

K-Nearest Neighbors

Distance metrics, the curse of dimensionality, choosing K, and its surprising link to vector search in modern LLMs.

Read more →
Day 7 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 25, 2026 · 10 min read

Naive Bayes

Bayes' theorem from scratch, the independence assumption, why it works despite being wrong, and text classification in production.

Read more →
Day 8 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 26, 2026 · 11 min read

Decision Trees

Gini impurity vs entropy, overfitting, pruning, feature importance — the foundation of every modern ensemble method.

Read more →
Day 9 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 27, 2026 · 11 min read

Bias, Variance & the Tradeoff

Bias and variance explained through darts, learning curves, and every model unified into one framework.

Read more →
Day 10 🌙 30-Day AI Ramadan Series
Deep Dive · Feb 28, 2026 · 11 min read

Regularization

L1, L2, Elastic Net, and Dropout — the tools that prevent overfitting in every model with intuition, math, and code.

Read more →
Day 11 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 1, 2026 · 12 min read

Bagging & Random Forests

How asking 500 imperfect models instead of one perfect model produces the most reliable predictions in classical ML.

Read more →
Day 12 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 2, 2026 · 13 min read

Boosting: AdaBoost & Gradient Boosting

Bagging makes models vote in parallel. Boosting makes them learn in sequence — each new model fixing exactly where the last one failed.

Read more →
Day 13 🌙 30-Day AI Ramadan Series
Practical Guide · Mar 3, 2026 · 13 min read

XGBoost & LightGBM

The engineering masterpieces that made gradient boosting 10–100× faster — second-order gradients, histogram splits, and early stopping.

Read more →
Day 14 🌙 30-Day AI Ramadan Series
Practical Guide · Mar 4, 2026 · 14 min read

Feature Engineering & Selection

The craft that separates good ML engineers from great ones — creating, transforming, and selecting features that make every model smarter.

Read more →
Day 15 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 5, 2026 · 23 min read

Neural Networks: From Logistic Regression to Deep Learning

A logistic regression unit, stacked and made non-linear — the bridge from classical ML to deep learning. Neurons, backprop, ReLU, and vanishing gradients from scratch.

Read more →
Day 16 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 6, 2026 · 15 min read

CNNs from Scratch

How machines learned to see — convolutions, filters, pooling, receptive fields, and residual connections explained from first principles.

Read more →
Day 17 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 7, 2026 · 14 min read

Transfer Learning & Fine-Tuning

How a model pre-trained on everyday photos achieved 95% accuracy on brain tumour classification with just hundreds of examples.

Read more →
Day 18 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 8, 2026 · 13 min read

Object Detection, YOLO & Segmentation

Classification tells you what. Detection tells you where. Segmentation tells you which exact pixels. YOLO, IoU, NMS, and U-Net explained.

Read more →
Day 19 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 9, 2026 · 13 min read

NLP Fundamentals

Teaching machines to read — tokenisation, bag-of-words, TF-IDF, and the foundational techniques that turn language into computable numbers.

Read more →
Day 20 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 10, 2026 · 29 min read

Recurrent Neural Networks

The first serious attempt at sequence modelling — how RNNs process text word by word, why gradients vanish over long sequences, and the exact mathematical proof that led to the LSTM.

Read more →
Day 21 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 11, 2026 · 35 min read

GRU & LSTM: Teaching Networks to Remember

Forget gates, cell states, and the gradient highway — how LSTM and GRU solved the vanishing gradient problem and powered a decade of NLP breakthroughs.

Read more →
Day 22 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 12, 2026 · 20 min read

Tokenization: How Text Becomes Numbers

BPE, WordPiece, SentencePiece — how GPT, BERT, and LLaMA split text into tokens, why the choice matters more than you'd think, and how to build a tokenizer from scratch.

Read more →
Day 23 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 13, 2026 · 35 min read

Word Embeddings: Word2Vec, GloVe & Sentence Transformers

king − man + woman ≈ queen. Word2Vec, GloVe, and sentence transformers — the geometry of meaning that powers semantic search and RAG retrieval.

Read more →
Day 24 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 14, 2026 · 35 min read

Attention & the Transformer Architecture

The architecture that powers GPT-4, Claude, and Gemini — self-attention, Q/K/V, multi-head attention, and the full transformer block built from scratch.

Read more →
Day 25 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 15, 2026 · 28 min read

Positional Encodings: Sinusoidal, RoPE & ALiBi

How transformers understand word order without recurrence — sinusoidal encodings, rotary position embeddings, and the techniques powering long-context models.

Read more →
Day 26 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 16, 2026 · 25 min read

BERT vs GPT: Encoders, Decoders & What They're For

Why BERT reads the whole sentence but GPT only looks left — masked language modelling vs causal language modelling, and when to use each architecture.

Read more →
Day 27 🌙 30-Day AI Ramadan Series
Deep Dive · Mar 17, 2026 · 22 min read

Tokenization Deep Dive: BPE, WordPiece & SentencePiece

A deeper look at how each major tokenizer algorithm works under the hood — merge rules, vocabulary construction, and why the same word tokenizes differently across models.

Read more →
Day 28 🌙 30-Day AI Ramadan Series
Practical Guide · Mar 18, 2026 · 26 min read

Fine-Tuning LLMs: Full, LoRA & QLoRA

Fine-tuning a 7B model on a single GPU — full fine-tuning, LoRA's low-rank decomposition, 4-bit QLoRA quantisation, and when each approach is the right call.

Read more →
Day 29 🌙 30-Day AI Ramadan Series
Practical Guide · Mar 19, 2026 · 28 min read

RAG: Retrieval-Augmented Generation Pipeline

Giving LLMs long-term memory — the full RAG pipeline from document ingestion and chunking to embedding, vector search, and grounded generation.

Read more →
Day 30 🌙 30-Day AI Ramadan Series
Practical Guide · Mar 20, 2026 · 25 min read

Prompt Engineering: Zero-Shot to Injection

Zero-shot, few-shot, chain-of-thought, system prompts, and prompt injection attacks — the complete picture of how to talk to language models and how they can be misled.

Read more →
All Blogs → All Categories
// career journey

Experience

Building real AI systems that make an impact.

May 2025 – Feb 2026

Freelance AI Engineer

Deep Learning, Computer Vision, NLP, Medical AI, MLOps

End-to-end AI system development for medical imaging and NLP applications. Built production-grade ML pipelines and deployed RAG systems.

Fine-tuned Qwen 2.5 7B VLM 4-bit Quantization RAG with FAISS Docker Deployment
Aug 2024 – Sep 2024

AI Research Intern

Neuronetix

Built XGBoost + Random Forest ensemble on patient data. Ranked Top 3 of 150 in customer churn prediction challenge.

AUC 0.92 F1 Score 0.89 Top 3/150
Mar 2024 – Jun 2024

Machine Learning Intern

ShAI

Built tabular music genre classifier using XGBoost. Built price prediction models with hyperparameter tuning via Optuna.

95% Accuracy 8 Classes MAE < $50
// learning

Education & Programs

🎓

B.Sc. Computer Science

Zagazig University 2019 – 2023

Coursework: Machine Learning, Deep Learning, NLP, Computer Vision, Algorithms, Databases. Graduation Project: 21 DL models for renewable energy forecasting (<15% MAPE).

💻

DEPI - ML Engineer

Digital Egypt Pioneers Initiative (Microsoft) Apr 2024 – Oct 2024

6-month intensive program (300+ hours) covering production ML, MLOps, and Azure deployment. Capstone: Breast cancer classifier achieving 97% accuracy on 15,000+ images.

Production ML MLOps Azure
🏆

Bootcamps & Training

IEEE, GDSC, DeepLearning.AI

Various technical bootcamps and certifications including IEEE technical programs, Google Developer Student Clubs, and Andrew Ng's Deep Learning Specialization.

IEEE Bootcamp GDSC Bootcamp DeepLearning.AI
// technical stack

Skills

LLMs & NLP

Qwen 2.5 LLaMA BERT GPT LoRA/QLoRA RAG LangChain FAISS Chroma Prompt Engineering

Computer Vision

Vision-Language Models CNNs DenseNet EfficientNet ResNet YOLO OpenCV MediaPipe Quantization

MLOps & Deployment

Docker FastAPI Flask MLflow Azure AWS Railway Cloudflare Workers CI/CD

Tools & Data

PyTorch TensorFlow Hugging Face Scikit-learn XGBoost LightGBM Pandas NumPy SQL Streamlit
⟩⟩ 100+ LeetCode problems solved • Python (Expert) • JavaScript, Java, SQL, Bash
// available for select projects

Have an AI problem worth solving? Let's ship it.

I partner with teams to design, fine-tune, and deploy production-grade ML systems — from RAG pipelines and LLM fine-tuning to medical computer vision. If you're exploring a collaboration, role, or technical advisory engagement, I'd love to hear about it.

Open to opportunities ⏱ Replies within 24h 🌍 Remote-friendly
✉ Start a conversation ↓ Download CV ⊕ Connect on LinkedIn

Let's Connect

Open to collaborations, opportunities, and interesting conversations about AI.

Email
Email
Mohameed.Abdalkadeer@gmail.com
→
LinkedIn
LinkedIn
Mo-Abdalkader
→
GitHub
GitHub
Mo-Abdalkader
→