SPARKs is a personal entrepreneurship project, awarded at both national and international levels. It aims to deliver innovative chemometrics-driven solutions and services at the intersection of data science and applied engineering.

This project is part of my chemometrics engineering thesis at the Université de Montréal. I worked on developing a self-driving laboratory platform to automate the inline optimization of plasmonic nanomaterial synthesis, integrating flow chemistry, inline UV-Vis spectroscopy, chemometrics, and machine learning. The approach connects autonomous synthesis to SERS-based analytical performance and extends toward neurochemical sensing applications, with relevance to neuroscience and other sectors requiring advanced molecular detection.

I had the honor to work on the EU-funded BarleyMicroBreed project, I led end-to-end high-throughput phenotyping to develop drought-tolerant barley varieties, integrating RGB, LiDAR, and hyperspectral data from ICARDA’s PhenoBuggy platform. Five innovative sub-projects : green cover segmentation, spike detection, unsupervised multimodal analysis, biomass prediction, and a custom phenotyping interface formed a robust ecosystem for genotype selection in arid regions. Leveraging advanced chemometric and deep learning methods, I achieved outstanding results (e.g., R=0.987, mAP50=0.943, R²=0.817), delivering scalable tools for precision agriculture and food security. This work showcases my expertise in data engineering, advanced analytics, and user-focused software development.

As a Chemometrician Agri-Data Scientist intern at GSMI from June to September 2025, I led the development and optimization of a technosoil amendment strategy to rehabilitate the Benguerir mine under a circular-economy framework. Using a RSM DOE (Design of Experiments) approach, I systematically screened organic (local composts, olive pomace) and mineral (e.g., clay, coal waste) amendments. Physicochemical analyses (XRF, FTIR, ICP, SEM) characterized both soil and amendment properties, while supervised and unsupervised chemometric models unraveled the key drivers of amendment efficacy. The resulting data-driven recommendations optimized amendment mixtures for enhanced soil structure, nutrient availability, and contaminant binding, laying the groundwork for sustainable land rehabilitation.

In collaboration with OCP Group and CBS-UM6P, I contributed to modeling and optimizing the factors affecting phosphoric acid slurry storage and filtration using the Six Sigma DMAIC methodology. I implemented a factorial DOE and supported filtration trials under controlled storage conditions, combined with chemical characterization of the filtered acid and solid residues. Multivariate analysis and predictive modeling were used to link storage parameters to key quality indicators and deliver actionable recommendations to improve process stability, filtration efficiency and phosphoric acid recovery.

In this project with OCP Group, I developed a high-performance analytical method for quantifying chlorides in the complex matrix of phosphoric acid. The approach began with the use of Design of Experiments (DOE) to systematically optimize analysis conditions. This was followed by a comprehensive analytical validation process, ensuring precision, accuracy, sensitivity, and compliance with standard protocols. To further enhance the method, I applied machine learning algorithms to model the data, improving both interpretation and predictive capabilities. This integration of experimental design, rigorous validation, and advanced data modeling resulted in a robust, efficient, and reliable method, well-suited for industrial applications.

During my internship at Pfizer, I worked on the validation of an HPLC method for quantifying sildenafil in pharmaceutical products, focusing on accuracy, precision, and regulatory compliance. I gained hands-on experience with analytical instruments (HPLC, GC, UV, IR) and participated in pharmaceutical quality control tests (dissolution, disintegration, physical tablet tests). I also observed microbiological testing, cleaning validation, and manufacturing processes, gaining a comprehensive view of pharmaceutical production and quality assurance.

As a chemometric engineer, I collaborated with agro-food engineers to develop and optimize a sweet potato-based yogurt. I applied chemometric tools (PCA, PLSR, DoE) for soil spectroscopy and spectral data modeling, identifying optimal conditions for sweet potato cultivation, as well as for milk quality assessment and production optimization. We enhanced product quality through statistical process control (SPC) and real-time monitoring using Power BI. This interdisciplinary project combined data-driven modeling with agro-food expertise to deliver a scalable, nutritious yogurt product.

In this academic project, a Python-based tool was developed to automate the validation of analytical methods for pharmaceutical industries, helping them ensure regulatory compliance and method reliability. It features a user-friendly graphical interface that allows users to upload data, select validation types (linearity, accuracy, repeatability, robustness, exactness...), and run tests without coding. The entire process is automated, generating detailed Word reports with statistical analysis and visualizations. This solution simplifies method validation, making it accessible and efficient for quality control and R&D teams.

As part of an academic project, I developed activated carbon from pineapple waste through a structured Design of Experiments (DoE) approach to optimize activation parameters. The material’s adsorption capacity was evaluated using methylene blue dye, with performance assessed via UV-Vis spectroscopy and modeled using adsorption isotherms and kinetics. Chemometric techniques were employed for data interpretation, and the results were benchmarked against a commercial adsorbent. This project highlighted the potential of agricultural waste valorization for sustainable water purification.

My portfolio is a dynamic and interactive website built with Next.js, Tailwind CSS, and Three.js to create an engaging, modern design with smooth animations using Framer Motion. It showcases my projects, skills, and achievements, allowing for personalized user interactions.