|Hi, I'm Malak Zaidi
Moroccan AI/ML Engineer & Full-Stack Developer | Learning and building every day
MSc student passionate about Machine Learning, LLMs, and real-world AI. Currently working on personal projects in MLOps, multi-agent systems, and fine-tuning open-source models.
</>About Me
Passionate AI researcher and software engineer with expertise in machine learning, distributed systems, and sustainable technology solutions
My Journey
As a highly motivated AI Master's student specializing in Artificial Intelligence and Distributed Systems, I bring a unique combination of MLOps pipelines, and scalable system architecture.
Currently based in Casablanca, Morocco, I'm actively researching advanced AI applications while focusing on green technology and sustainable AI solutions.
Key Achievements
- 95% accuracy in healthcare AI model development
- 40% reduction in clinical reporting time through automation
- 7% improvement in LLM code generation performance
- Production-ready MLOps pipelines with Docker & Airflow
AI Innovation
Cutting-edge ML/DL solutions
Results-Driven
High-performance AI systems
Team Leadership
Strong communication skills
Rapid Learning
Quick tech adaptation
Featured Projects
Production-ready AI/ML solutions with measurable business impact and technical excellence
AI & Machine Learning Solutions
Revolutionary multi-agent RAG system automating structured MRI radiology reports from voice/text input
5 specialized agents (Classification → Extraction → Structuring → Validation → Generation)
Local DeepSeek-R1 1.5B + ClinicalBERT for privacy-first processing
92% speech recognition • 95% semantic validation • <30s generation time
GDPR/HIPAA compliant • Full audit trail • Deployed in clinical research
Fine-tuned LLaMA-3.2-1B-Instruct into a high-performance code assistant using QLoRA + Unsloth
Achieved 22.0% pass@1 on HumanEval (+57% over base)
Trained in 45 minutes on free Google Colab T4 GPU
45.1M trainable params (3.52%) • 4-bit NF4 quantization • 8k Python tasks
Outperforms 7B models in resource-constrained environments
Production-ready sales forecasting pipeline powered by Astro (Apache Airflow)
Fully orchestrated ETL → training → inference → monitoring
Dynamic task mapping • MLflow tracking • MinIO artifact storage
Dockerized • CI/CD ready • Scalable across cloud or on-prem
Built for enterprise: zero downtime deployments, audit trails, alerts
Transforming education at ENSET Mohammedia with an intelligent RAG-powered tutor
Powered by LLaMA 3.2 + PostgreSQL pgvector for curriculum-aligned answers
Built with Java Spring Boot + Spring AI • Docker • JWT auth
Real-time, context-aware responses from verified course materials
Live deployment • 500+ active students • Admin analytics dashboard
Full-stack educational data science project analyzing 480 student records
Random Forest model predicts performance with 92% accuracy
Identified 24% at-risk students using engagement, absence, parental factors
Interactive Power BI dashboard with DAX, filters, and early warning system
Delivered actionable insights to ENSET faculty for targeted interventions
Data & Machine Learning Labs
ChartSupervised Learning Projects
99.94% accuracy with PCA + GridSearchCV
GamepadReinforcement Learning Projects
DocumentTechnical Whitepapers
Formal mathematical framework for autonomous AI agents in distributed environments. Bridges decision theory, reinforcement learning, and distributed systems with rigorous analysis of agent behavior, convergence properties, and performance guarantees.
Comprehensive exploration of Transformer architecture using an intuitive library research team analogy. Covers mathematical foundations, self-attention mechanism, positional encoding, and multi-head attention with applications in NLP and computer vision.
Technical whitepaper on optimizing AI models for edge devices through quantization. Addresses deployment challenges between model complexity and hardware constraints with practical strategies for mobile, IoT, and robotics applications.
Technical whitepaper exploring Deep Q-Networks (DQN) and Double DQN agents. Covers Q-learning foundations, experience replay, target networks, overestimation bias reduction, Python implementations, performance comparisons, and applications in gaming, robotics, and autonomous systems.
Full-Stack Applications
Climate solutions platform for African environmental initiatives
Microservices architecture with Spring Cloud & Netflix Eureka
Enterprise security: JWT authentication & role-based access
Modern frontend: React.js with Material UI components
Academic recognition: Praised for innovation & technical excellence
Full-stack banking platform for account management, transactions, and analytics
Supports current/savings accounts, debit/credit/transfer operations
Enterprise-grade security via Spring Security + JWT • Role-based access
Interactive Angular dashboard with Chart.js • Mobile-responsive UI
MySQL backend with JPA/Hibernate • Swagger docs • Production-ready
Blockchain Applications
Private blockchain for organ donation traceability & transplant coordination
Advanced security: RSA-2048 authentication & SHA-256 hashing
Adaptive PoW consensus: Priority-based mining for critical cases
Smart matching engine: Compatibility scoring (blood, HLA, urgency)
Real-time dashboard: Flask web interface with Tailwind CSS
Featured Projects
Production-ready AI/ML solutions with measurable business impact and technical excellence
🤖 AI & Machine Learning Solutions
Revolutionary multi-agent RAG system automating structured MRI radiology reports from voice/text input
5 specialized agents (Classification → Extraction → Structuring → Validation → Generation)
Local DeepSeek-R1 1.5B + ClinicalBERT for privacy-first processing
92% speech recognition • 95% semantic validation • <30s generation time
GDPR/HIPAA compliant • Full audit trail • Deployed in clinical research
Fine-tuned LLaMA-3.2-1B-Instruct into a high-performance code assistant using QLoRA + Unsloth
Achieved 22.0% pass@1 on HumanEval (+57% over base)
Trained in 45 minutes on free Google Colab T4 GPU
45.1M trainable params (3.52%) • 4-bit NF4 quantization • 8k Python tasks
Outperforms 7B models in resource-constrained environments
Production-ready sales forecasting pipeline powered by Astro (Apache Airflow)
Fully orchestrated ETL → training → inference → monitoring
Dynamic task mapping • MLflow tracking • MinIO artifact storage
Dockerized • CI/CD ready • Scalable across cloud or on-prem
Built for enterprise: zero downtime deployments, audit trails, alerts
Transforming education at ENSET Mohammedia with an intelligent RAG-powered tutor
Powered by LLaMA 3.2 + PostgreSQL pgvector for curriculum-aligned answers
Built with Java Spring Boot + Spring AI • Docker • JWT auth
Real-time, context-aware responses from verified course materials
Live deployment • 500+ active students • Admin analytics dashboard
Full-stack educational data science project analyzing 480 student records
Random Forest model predicts performance with 92% accuracy
Identified 24% at-risk students using engagement, absence, parental factors
Interactive Power BI dashboard with DAX, filters, and early warning system
Delivered actionable insights to ENSET faculty for targeted interventions
🔬 Data & Machine Learning Labs
ChartSupervised Learning Projects
Comprehensive KNN implementation on the Agaricus-Lepiota dataset with feature engineering, PCA visualization, and hyperparameter tuning — achieving 99.94% accuracy using K=1 + Manhattan distance.
Key Highlights
- →Correlation heatmap + feature importance analysis
- →2D/3D PCA projections show perfect class separation
- →GridSearchCV over K, distance metric, and weights
Advanced SVM with RBF kernel on raw CIFAR-10 pixels — 42.8% accuracy without CNNs. Full support vector analysis, PCA decision boundaries, and kernel comparison.
Key Highlights
- →RBF kernel beats linear by +15.3%
- →97% of training data used as support vectors
- →2D PCA visualization of 10-class decision boundaries
More supervised learning implementations including regression analysis, clustering algorithms, and advanced feature engineering techniques.
🎮Reinforcement Learning Projects
Comprehensive RL framework using Open AI Gymnasium to train agents that navigate a 5×5 grid, pick up passengers, and optimize delivery routes—achieving reward improvement from -800 to +7.28 in 500 episodes.
Key Highlights
- →3 algorithms (Q-Learning, SARSA, DQN) with epsilon-greedy & experience replay
- →Interactive Plotly visualizations with policy maps & optimal path traces
- →Cross-environment validation (Taxi-v3, FrozenLake, CartPole)
Implemented both standard DQN and Double DQN agents using TensorFlow/Keras to navigate a 4×4 grid world, avoiding obstacles and reaching goal positions through experience replay and target network optimization.
Technical Architecture
- →2-hidden layer MLP (24 neurons each) with ReLU activation for Q-value approximation
- →Experience replay buffer (2000 capacity) with batch learning (size 32)
- →Double DQN reduces overestimation bias with separate target network
📄Technical Whitepapers
Formal mathematical framework for autonomous AI agents in distributed environments. Bridges decision theory, reinforcement learning, and distributed systems with rigorous analysis of agent behavior, convergence properties, and performance guarantees.
Comprehensive exploration of Transformer architecture using an intuitive library research team analogy. Covers mathematical foundations, self-attention mechanism, positional encoding, and multi-head attention with applications in NLP and computer vision.
Technical whitepaper on optimizing AI models for edge devices through quantization. Addresses deployment challenges between model complexity and hardware constraints with practical strategies for mobile, IoT, and robotics applications.
Technical whitepaper exploring Deep Q-Networks (DQN) and Double DQN agents. Covers Q-learning foundations, experience replay, target networks, overestimation bias reduction, Python implementations, performance comparisons, and applications in gaming, robotics, and autonomous systems.
💻 Full-Stack Applications
🌍 Climate solutions platform for African environmental initiatives
🏗️ Microservices architecture with Spring Cloud & Netflix Eureka
🔐 Enterprise security: JWT authentication & role-based access
⚛️ Modern frontend: React.js with Material UI components
🎯 Academic recognition: Praised for innovation & technical excellence
Comprehensive full-stack banking platform for account management, transactions, and analytics
Supports current/savings accounts, debit/credit/transfer operations with real-time updates
Enterprise-grade security via Spring Security + JWT • Role-based access • Audit trails
Interactive Angular dashboard with Chart.js visualizations • Mobile-responsive UI
MySQL backend with JPA/Hibernate • Swagger docs • Production-ready with CI/CD
⛓️ Blockchain Applications
🏥 Private blockchain for organ donation traceability & transplant coordination
🔐 Advanced security: RSA-2048 authentication & SHA-256 hashing
⚡ Adaptive PoW consensus: Priority-based mining for critical cases
🎯 Smart matching engine: Compatibility scoring (blood, HLA, urgency)
📊 Real-time dashboard: Flask web interface with Tailwind CSS
Professional Experience
Hands-on experience in AI/ML research, data science, and full-stack development
Data Science Trainee
WorldQuant University Applied Data Science Lab
Hands-on data science experience with real-world datasets and industry-standard tools. Executing end-to-end ML projects, API integration, and advanced statistical analysis. Key focus: Production-ready data pipelines and business intelligence solutions.
Education Journey
Master in Distributed Systems & Artificial Intelligence (SDIA)
CurrentENSET Mohammedia – University Hassan II
Sep 2024 – Jun 2026 (Ongoing)
Research-Grade AI & Distributed Computing:
- ▹ Transformers, LLMs, Generative & Agentic AI (CrewAI, RAG, XAI)
- ▹ Advanced Deep Learning: CNN, RNN, Edge AI, Transfer Learning
- ▹ NLP, Knowledge Graphs, Semantic Web, Monte Carlo Methods
- ▹ MLOps, Big Data, Airflow, MLflow, GPU Clusters
- ▹ Blockchain, Cryptography, Parallel & Distributed Systems
B.Sc. Mathematical & Computer Sciences
FS Ben M’Sik – Specialty: Databases & Information Systems
Sep 2021 – Jun 2024 • Magna Cum Laude
Full-Stack Systems & Math Engineering:
- ▹ Advanced Java, Spring Boot, Microservices, JPA/Hibernate
- ▹ C, x86 Assembly, Linux Kernel, Operating Systems
- ▹ Deep Math: Linear Algebra, Numerical Analysis, Probability & Statistics
- ▹ Database Engineering (SQL/NoSQL), Performance Optimization
- ▹ Algorithm Design, Data Structures, Computational Complexity
Baccalauréat – Mathematical Sciences (Option B – Technological)
Ibn Mandour High School, Casablanca
Sep 2020 – Jun 2021
Morocco’s Elite Engineering-Prep Track:
- ▹ Advanced Mathematics & Engineering Sciences
- ▹ Arduino Robotics, Embedded Systems, Digital Electronics
- ▹ Algorithms, Basic Programming, Systems Thinking
- ▹ Physics, Mechanics, Automation, Signal Processing
- ▹ National Olympiad-level complex problem solving
Option B = Morocco’s STEM track for future engineers (vs Option A = Medicine)
Technical Expertise
Comprehensive technical stack spanning AI/ML, cloud computing, and full-stack development
AI & Machine Learning
MLOps & Production
Full-Stack Development
Data Science & Analytics
Leadership & Communication
Certifications & Achievements
Industry-recognized certifications spanning AI/ML, Cloud Computing, Data Management, and Professional Development
Artificial Intelligence & Machine Learning
Comprehensive 4-course specialization covering regression, classification, clustering, and information retrieval. Gained hands-on experience building intelligent applications using real-world datasets and modern ML frameworks.
Completed practical simulation involving data extraction, analysis, and development of an AI-powered financial chatbot. Demonstrated proficiency in applying generative AI to real-world business challenges.
Cloud Computing & Infrastructure
Foundational certification covering cloud service models (IaaS, PaaS, SaaS), deployment models, and core cloud concepts. Understanding of cloud architecture and enterprise cloud solutions.
Blockchain certification (LFS170) covering distributed ledger technology, use cases, and practical implementations. Understanding of blockchain architecture and its applications across industries.
Data Management & Analytics
Enterprise-level certification in database systems, data warehousing, and analysis techniques. Expertise in managing large-scale data infrastructure and implementing data governance strategies.
Sustainable Technology
24-hour certification in sustainable AI practices and green digital technologies. Focused on implementing eco-friendly technology solutions and reducing environmental impact through digital transformation.
Professional Development & Soft Skills
Certification in strategic thinking frameworks and systematic problem-solving methodologies. Acquired skills in critical analysis, decision-making, and innovative solution development.
Professional training in effective communication, presentation skills, and audience engagement. Developed confidence in delivering impactful presentations and leading technical discussions.
Languages & Communication
Multilingual fluency enabling global collaboration and cross-cultural leadership
Native speaker with professional writing and presentation skills
Fluent in technical communication and international collaboration
Native speaker enabling MENA region project leadership
Professional working proficiency for business communication
Basic conversational skills for international tech collaboration
Why Choose Me?
Unique fusion of **technical mastery**, **research innovation**, and **global vision** — delivering impact that lasts.
Innovation-Driven Problem Solver
I don’t just implement — I **invent**. My multi-agent healthcare AI achieved **95% accuracy** by orchestrating specialized models, proving I think beyond boundaries to deliver **breakthrough results**.
Global Perspective & Cultural Intelligence
Fluent in **5 languages**, I bridge cultures and markets. This enables me to lead **international teams**, understand **diverse user needs**, and communicate complex AI concepts **across borders**.
Proven Track Record of Impact
**40% faster** medical reporting. **22% better** code generation. **Production-grade MLOps**. My work doesn’t just run — it **transforms businesses**.
What Sets Me Apart
Extracurricular & Activities
Leadership, community impact, and professional growth — **beyond the classroom**
Active contributor with **Pull Shark** and **Quickdraw** badges. Mentoring junior devs and driving community AI/ML projects.
Regular speaker at university seminars and conferences — sharing **AI research**, **sustainable tech**, and **real-world impact**.
Leading **Green AI** initiatives to reduce environmental impact of AI systems and promote **responsible tech development**.
Let's Connect
Ready to **redefine AI/ML innovation**? Let’s explore how my expertise can **elevate your team**.
Send me a message
Get in touch
I'm actively seeking AI/ML and Data Science internship opportunities. Open to innovative projects, research collaborations, and contributing to your team's technical goals.
Location
Casablanca, Morocco • Open to Remote
Resources & Publications
AI/ML & Data Science internships
Ready to contribute to innovative projects
Open to remote & international roles
Available for interviews & assessments