DOGBALOU MOTOGNON WASTALAS D'ASSISE
Full-Stack Web Developer | Applied AI Research | AI Technical Solutions
Embedded Files
About Me
Hello! I'm Wastalas, a PhD researcher in Artificial Intelligence and a full-stack developer passionate about building intelligent systems and scalable web applications. I work at the intersection of data science, machine learning, and software engineering, integrating AI, LLM, and ML techniques to solve real-world problems. I am fluent in both English and French and enjoy developing solutions that transform complex data into meaningful insights.
Download my CV through the following button:(Updated March 2026)
My Projects
STACK-AI-Math-Course (PhD Project)
STACK-AI-Math-Course (PhD Project)
STACK-AI-Math-Course is an advanced, adaptive learning platform for university-level mathematics, currently focused on Linear Algebra and Differential Calculus. It serves as an intelligent, always-available tutor that uses generative AI to provide reasoning-based grading, scanned solution analysis, and detailed feedback on students’ mathematical logic and problem-solving processes, going far beyond traditional answer checking.
The platform allows students to submit solutions directly either through LaTeX-based input or scanned handwritten work and receive AI-generated feedback, hints, and step-by-step solutions. It continuously tracks student progress to generate personalized practice problems, supports exam creation and scheduling, provides student dashboards, and delivers results by email, offering a scalable, mobile-first solution for modern mathematics education.
Technology Stack
Frontend
The frontend is built using Next.js 15 with the App Router, leveraging React 19 and TypeScript for a robust and maintainable architecture. Styling and UI components are implemented with Tailwind CSS v4, shadcn/ui based on Radix Primitives, Lucide React icons, and utility helpers such as clsx and tailwind-merge. Mathematical notation is rendered using KaTeX and react-katex, ensuring fast and accurate display of LaTeX mathematics.
Backend and Infrastructure
The backend is powered by Convex, providing a real-time database and backend-as-a-service architecture. Authentication is handled through Clerk, and the platform runs on Node.js and Bun for efficient server-side execution.
Artificial Intelligence
The AI layer is designed with a multi-model architecture, allowing the system to switch between models based on task complexity, performance, and cost considerations. Current integrations include OpenAI GPT-4o and GPT-4o-mini via REST APIs, as well as Google Gemini 1.5 Flash using the official SDK. This flexible design ensures high-quality reasoning feedback while maintaining scalability and efficiency.
STACKAI-Math-Educator (PhD Research Project)
STACKAI-Math-Educator (PhD Research Project)
STACKAI-Math-Educator is a full-stack analytics and adaptive feedba ck platform designed to support university lecturers using the STACK Assessment system. The platform enables educators to ingest, process, and analyze student responses at scale, transforming raw assessment data into interpretable learning diagnostics.
By combining advanced data visualization with AI-driven pedagogical analysis, the system provides insights into common misconceptions, reasoning errors, and at-risk students, moving beyond traditional grading toward meaningful formative assessment.
A core feature of the platform is the automated application of Newman’s Error Analysis, powered by large language models and retrieval-augmented generation (RAG). This allows lecturers to identify recurring reasoning patterns, conceptual gaps, and feedback categories with high precision, supporting evidence-based instructional decisions. The platform is deployed as a secure, serverless web application, offering scalable access for academic staff.
Technology Stack
Frontend
Frontend
The frontend is built with Next.js 16 and TypeScript, providing a modern, performant, and maintainable React-based architecture. The interface is designed to support data exploration and visualization workflows for lecturers, with a focus on clarity, responsiveness, and usability.
Backend and Infrastructure
Backend and Infrastructure
The backend leverages Convex as a real-time database and backend-as-a-service solution, enabling reactive data flows and simplified serverless architecture. Authentication and access control are managed using Clerk, ensuring secure and role-based access for academic users. The application is deployed on Vercel and Convex Cloud, allowing for scalable, low-latency global access.
Artificial Intelligence
Artificial Intelligence
The AI layer integrates OpenAI GPT-4o with retrieval-augmented generation (RAG) pipelines to analyze imported STACK data. This architecture supports automated reasoning analysis, classification of student errors, and the generation of pedagogically meaningful feedback, bridging quantitative assessment results with qualitative educational insight.
Student Error Explainer (PhD Research)
Description: A full-stack intelligent platform that automatically analyzes, anonymizes, and categorizes student mathematical errors using AI. Designed for educators and researchers to understand misconceptions and provide targeted pedagogical feedback at scale.
Key Features:
Secure data anonymization pipeline (CSV/XLSX) — strips PII while preserving data integrity
AI-powered error explanation using GPT-4o with real-time feedback generation
Automated error categorization using Newman Error Theory (5-stage classification framework)
Asynchronous job processing for scalable analysis without API timeouts
Role-based access control via Clerk authentication
Tech Stack:
Frontend & API:
Framework: Next.js 16.1.6, React 19.2.3
Language: TypeScript
Styling: Tailwind CSS 4
Authentication: Clerk
HTTP Client: Fetch API
Backend & Data Processing:
Runtime: Node.js with TypeScript
Database: MongoDB Atlas (cloud)
AI/LLM: OpenAI GPT-4o
File Processing: XLSX library (Excel/CSV parsing)
Job Processing: Custom polling queue (MongoDB-backed)
Infrastructure & Deployment:
Frontend & Serverless APIs: Vercel
Background Worker: Railway
Project Title: Personal Day's Tasks (React Native)
Description: A full-stack React Native productivity app featuring real-time synchronization, smart task prioritization, and daily focus management. Personal Day's Tasks is a mobile "Focus Manager" designed to help to organize daily workflow efficiently. Unlike standard to-do lists, it centers on a "Home Dashboard" that intelligently highlights the most critical high-priority task of the day. The application supports Time Blocking strategies by allowing to schedule start times and durations for tasks. It utilizes a priority system (High/Medium/Low) to sort and surface the most important items automatically.
Built with offline-first principles and real-time data capabilities, changes sync instantly across all devices. The UI features a custom dark-mode-ready design system with smooth gradients and interactive animations to encourage productivity.
Tech Stack:
Mobile Framework: React Native (Expo SDK 54)
Language: TypeScript
Backend & Database: Convex (Real-time, Serverless)
Routing: Expo Router
Styling: Custom Component System with expo-linear-gradient
Icons: Ionicons
Key Features:
Real-Time Sync: Instant data updates via Convex WebSocket connection.
Smart Dashboard: Dynamic "Current Focus" card that adapts to priorities.
Time Management: Integrated time-blocking tools for setting task durations.
Cross-Platform: Optimized for both iOS and Android.
Live:
Project: React Admin Dashboard
Description: A modern, admin dashboard application built with React and Vite. This project demonstrates a responsive, data-driven interface designed for scalability and user experience. It features a sleek dark-mode aesthetic utilizing Tailwind CSS and incorporates fluid animations with Framer Motion to enhance interactivity.
The dashboard provides real-time data visualization through Recharts, offering intuitive insights into sales, user demographics, and product performance. It includes a fully functional client-side routing system using React Router and integrates with the GitHub API.
Tech Stack
Frontend: React.js (v18), Vite
Styling: Tailwind CSS, PostCSS
Animations: Framer Motion
Navigation: React Router DOM
Data Visualization: Recharts
Icons: Lucide React
Deployment: Vercel
Key Features
Interactive Data Visualization: Dynamic line, bar, and pie charts for detailed analytics (Sales, Orders, Users).
Responsive Design: Fully responsive sidebar and layout that adapts seamlessly to desktop and mobile devices.
Modern UI/UX: Glassmorphism effects, smooth transitions, and a polished dark theme.
Modular Architecture: Component-based structure ensuring code reusability and maintainability.
Live: https://react-admin-dashboard-seven-alpha.vercel.app/
Python–JavaScript Chat App
Python–JavaScript Chat App
Description: A real-time chat application built with Flask and Flask-SocketIO, featuring live WebSocket communication and deployed on Render using Gunicorn and Eventlet.
Tech: Python • Flask • Flask-SocketIO • JavaScript • Gunicorn • Eventlet • Render
Live: https://python-javascript-chatapp.onrender.com
Interactive Math Quiz System
Description: I developed an interactive mathematics assessment platform that integrates AI-generated feedback with student written reasoning and LaTeX support. The system features adaptive hints, step-by-step solutions, instant correctness checking, and dynamic question generation. Developed through continuous student and lecturer feedback.
Tech: Node.js • Express • Supabase • PostgreSQL • OpenAI API • JWT • HTML • CSS • JavaScript • MathJax • Render
Live: https://interactive-math-quizv2.onrender.com/
In progress
Page updated
Google Sites
Report abuse