AI engineer with a product mindset. I design, build, evaluate, and deploy machine-learning systems — from predictive models and computer vision to RAG pipelines, AI agents, and intelligent applications.
|
Machine Learning Predictive modeling · Feature engineering · Explainability · Evaluation Deep Learning PyTorch · TensorFlow · Neural networks · Computer vision Generative AI LLMs · RAG · Agentic workflows · Multi-agent systems · AI evaluation |
Backend Python · FastAPI · Flask · REST APIs · PostgreSQL AI Infrastructure LangChain · LangGraph · ChromaDB · Docker · CI/CD Product Rapid prototyping · Deployment · Automation · UX-aware AI products |
RAW DATA
│
▼
┌─────────────────┐
│ ENGINEER │
│ features / data │
└────────┬────────┘
│
▼
┌─────────────────┐
│ INTELLIGENCE │
│ ML / DL / LLMs │
└────────┬────────┘
│
▼
┌─────────────────┐
│ ORCHESTRATE │
│ RAG / Agents │
└────────┬────────┘
│
▼
┌─────────────────┐
│ SHIP IT │
│ API / App / Ops │
└─────────────────┘
A visual scientific-computing project exploring near-Earth-object trajectories, gravitational focusing, atmospheric entry, impact effects, and ML-assisted simulation.
Python Machine Learning Orbital Mechanics Scientific Computing Three.js WebGL
|
Wildfire early-warning & relief coordination for Algeria. Satellite-based fire tracking, fire-danger forecasting, and relief-oriented information in one system.
|
Computer vision for street-littering detection. Built around object detection, pose signals, and abandoned-object reasoning for evidence-oriented workflows.
|
|
An advanced RAG framework for multi-step reasoning, dynamic retrieval, and precise context-aware answers.
|
Turns natural language into entities + relationships and makes knowledge navigable through an interactive graph.
|
|
Healthcare prior-authorization intelligence built around local multi-agent RAG workflows.
|
A production-inspired evaluation project covering hallucination, factuality, toxicity, reasoning quality, bias, and semantic similarity.
|
|
Customer-churn prediction with engineered features, model explainability, and AI-powered recommendations.
0.919 AUC · 130+ engineered features |
AI study assistant that turns educational video content into structured notes, summaries, and quizzes.
|
| Typical ML project | How I approach it |
|---|---|
| Train a model | Build the complete workflow |
| Optimize a metric | Measure the right outcome |
| Notebook-first | Product + API + deployment |
| Prompt → answer | Retrieval → reasoning → evaluation |
| Demo once | Design for repeatable use |
| Model as the finish line | Model as one component of a system |
LangChain · LangGraph · RAG · ChromaDB · XGBoost · SHAP · NLP · Computer Vision · Agentic AI
|
AI AGENTS Tool use Planning Multi-step reasoning Agent evaluation |
PRODUCTION AI FastAPI Docker MLOps Monitoring |
SYSTEM DESIGN RAG architectures Knowledge systems Reliable AI workflows |
↳ Computer Vision
Trash Sentinel · Hand Gesture Recognition · Face Mask Detection · real-time OpenCV experimentation
↳ Applied Machine Learning
ChurnIQ · Gymshark Recommendation System · Student Performance Prediction · Handwritten Digit Classification
↳ LLM / NLP
Agentic RAG · Knowledge Graph RAG · AI Agent Evaluation · AI Job Search Agent · AI News PDF Summarizer · NoteStream
↳ Engineering / Automation
AutoGit · REST-to-MCP experimentation · Dockerized AI services · FastAPI systems
class AEK:
focus = [
"AI / Machine Learning Engineering",
"LLM Applications & Agentic Systems",
"RAG & Knowledge Systems",
"Computer Vision",
"Production-minded AI"
]
principle = "Build useful systems, not impressive notebooks."AI / ML Engineering · Applied AI · LLM Engineering · AI Automation · Remote Opportunities · Collaborations
