Bridging Biochemistry & Applied Statistics

• Collaborated with an inter-disciplinary team on developing an end-to-end modeling and scenario analysis for Bunge's shea supply chain, creating a digital twin and actionable planning guidelines.
• Delivered a prototype scenario computation tool in Dataiku enabling multi-level optimization and valorization.
• Identified strategies with potential annual savings of €1M+ including early kernel buying, origin optimization, and improved processing schedules.

Title: Assessment of Liver and Kidney Function, and Lipid Profile of Ntonso Traditional Textile Dyers

• Designed and executed a case-control study (n=100) to assess biochemical impacts of occupational exposure.
• Engaged local community stakeholders to explain study objectives and obtain informed consent, ensuring ethical compliance and participant trust.
• Investigated effects of heavy metal-containing dyes on liver and kidney function in textile workers.
• Performed biochemical assays (ALT, AST, ALP, creatinine, lipid profile) and interpreted clinical data.
• Applied statistical techniques (Student's t-test, Mann-Whitney, Pearson correlation) using SPSS.
• Analyzed multi-variable datasets to evaluate liver, kidney, and lipid biomarkers; interpreted p-values and confidence intervals.
• Synthesized findings into actionable recommendations for occupational health policy.

Bunge Limited

• Designed and standardized an inter-laboratory proficiency testing protocol.
• Applied advanced statistical methods (Mandel's h & k statistics, Grubb's outlier test, Z-scores, Youden plots) to evaluate precision, accuracy, and reproducibility across 13 laboratories (21 laboratories in 2025).
• Conducted data validation and performance benchmarking, identifying sources of variability and recommending corrective actions.
• Authored a comprehensive report and established standardized procedures for ongoing proficiency testing.

Bunge Limited

• Developed a rapid, chemical-free testing method to determine shea stearin quality using basic Refractive Index measurements.
• Replaced expensive and slow wet-chemistry analysis with a machine learning model, drastically reducing lab costs.
• Significantly improved turnaround time for quality screening, allowing for faster production decisions.