About
I am Jaouad Chakir, a software engineer and deep learning enthusiast based in Morocco. I specialize in backend development and building scalable, efficient systems using Python, TypeScript, and C++. My skills include developing RESTful APIs, working with frameworks like Django and NestJS, and deploying containerized applications with Docker.
I am also passionate about deep learning and enjoy solving real-world problems through data analysis, models fine-tuning, or even models selection. I have experience with neural networks, and other Deep Learning algorithms, using tools like NumPy, Keras, and PyTorch.
I am currently a student at 1337 Coding School (42 Network), where I focus on enhancing my skills in software engineering and deep learning. My GitHub projects showcase my work in backend systems, machine learning models, and algorithm development.
When I’m not coding, I explore advanced technologies to find innovative solutions to everyday challenges, or simply play football.
Education
1337 Coding School
2021 - 2025
Peer-to-peer, project-based curriculum with no formal instructors, covering C/C++ systems programming, algorithms, web development, and full-stack projects under strict deadlines and code-quality standards. Learned advanced problem-solving and collaboration skills through peer-based challenges.
BTS Al Khawarizmi
2017 - 2019
Two-year technical degree in network administration, Linux system deployment, and infrastructure security. Gained in-depth knowledge in Linux, network design, and CCNA 200-301 fundamentals.
Experience
Software Engineer, AI Solutions at Marwa Retail
Aug 2025 - Present
- Architected a scalable SaaS platform utilizing Express.js and FastAPI to manage inventory data structures.
- Engineered AI/deep learning algorithms for smart allocation, automated restocking, and stock transfers, optimizing supply chain logistics.
- Containerized services with Docker for consistent deployments across development and production environments.
Software & AI Intern at Sofrecom Morocco
Feb 2023 - Aug 2024
- Built scalable RESTful APIs with Django and optimized query performance.
- Integrated AI/ML models into applications to generate predictions.
- Collaborated within an Agile team using Git for version control and code reviews.
- Developed an automated tag prediction system using XGBoost in the Tagma project.
- Explored AI model training and tested ML algorithms and neural networks for better accuracy.
Projects: Software-Engineering
Transcendence: Online Ping-Pong Game and Chat
Real-time ping-pong game using WebSockets with secure chat, JWT authentication, and two-factor login to protect user sessions and matchmaking data.
Inception: Docker/Microservices Architecture
Robust microservices architecture with Docker and Nginx load balancing, running isolated MariaDB and WordPress services in dedicated containers with persistent volumes.
Containers: STL-like Container Implementation
Recreated STL containers like vector and map using C++ templates and Red-Black Trees.
WebServ: HTTP Server
Developed a reliable HTTP server using C++ sockets and Linux syscalls for connection management.
Projects: Machine-Learning
Tagma: Automated Tag Prediction System
Built a model to predict project tags and streamline classification, improving data quality for higher accuracy.
MLP-from-scratch: Multi-Layer Perceptron
Implemented a neural network from scratch for classification tasks, coding forward and backward propagation and gradient descent using only NumPy.
Tweets-NLP: NLP and Sentiment Analysis
Prepared tweet data to classify sentiment and measure similarity, employing multiple ML algorithms for comprehensive analysis.
Churn: Bank Data Processing and Models Training
Banking data preprocessing and training multiple Machine Learning models, including Naive Bayes, Random Forest, and MLP.
Product-Similarity: Product Category Matching
Fine-tuned a pretrained vision model with PyTorch using triplet loss to learn a metric embedding space that identifies product categories from images without retraining.
Visual-Search: Image-Based Product Retrieval
Vectorized product images with DINOv3 and FashionCLIP embeddings stored in a vector database, retrieving top-k similar products for unseen items via nearest-neighbor queries.
Technical Skills
Skills
Machine Learning Tools
Programming Languages
Frameworks
Tools & Technologies
Languages
Arabic
Native
English
Upper Intermediate
French
Intermediate