Overview
Develop a computational screening workflow combining physics-based modeling (QM/MD) and machine learning (QSPR/ML) to predict and rank natural sweetener candidates with:
- Higher sweetness potential
- Lower bitterness/off-taste risk
Responsibilities
Key Responsibilities:
Data Curation & Molecular Feature Generation (Month 1)
- Curate a dataset of known natural sweeteners and reference compounds
- Compute molecular descriptors and key physicochemical properties
- Prepare analysis-ready datasets for modeling and validation
Physics-Based Modeling for Mechanistic Insights (Months 2–5)
- Perform QM calculations to analyze interaction-related molecular properties
- Run targeted MD simulations to study binding/stability trends and support structure–taste hypotheses
- Summarize computational findings into interpretable insights
Machine Learning / QSPR Model Development (Months 2–5)
- Build predictive models linking molecular features to sweetness/bitterness outcomes
- Evaluate model performance using retrospective validation on known compounds
- Identify and communicate key molecular drivers influencing taste outcomes
Decision Support & Candidate Prioritization (Month 6)
- Integrate QM/MD + ML outputs into a reusable screening workflow
- Generate a ranked list of high-potential natural sweetener candidates
- Provide clear guidance for experimental or formulation follow-up
Scope:
In Scope:
- Computational analysis (QM/MD + ML)
- Retrospective validation on known compounds
Out of Scope:
- Large-scale sensory testing or production testing
Deliverables 🎯:
- Reusable modeling workflow (computational screening framework)
- Predictive ML models for sweetness vs bitterness risk
- Computational insights supporting structure–taste relationships (QM/MD-informed)
- Ranked list of high-potential natural sweetener candidates
- Final technical report + summary slide
- Knowledge transfer session to hand over methods, results, and recommendations
Qualifications
Title: Computational Modeling & Machine Learning Intern – Natural Sweetener Taste Profile Improvement (6 Months)
Duration: 6 months (Master’s internship)
Project/Domain: Natural sweetener development – computational screening for sweetness improvement and bitterness/off-taste risk reduction
Business Context:
Natural sweeteners support clean-label and sugar-reduction strategies, but many candidates suffer from bitterness, off-tastes, or delayed onset. Experimental optimization can be slow and resource-intensive. This internship focuses on generating actionable computational insights and a screening framework to help prioritize the most promising natural sweetener candidates before experimental testing.
