Amna Mnazar

I am a highly adaptable Data Scientist with an MSc in Artificial Intelligence & Robotics, specializing in turning complex, ambiguous datasets into high-impact, actionable business insights. My core strength is demonstrated through my consistent top-rankings on competitive data science platforms like Lunor Quest.

Core Expertise & Proven Achievements

My expertise spans the full lifecycle of Machine Learning and Advanced Analytics, applied across diverse, real-world domains:

Quantitative Finance & Time Series: Secured 2nd Place in the highly competitive FOMC Rate Prediction challenge by developing a robust multi-class ensemble model, integrating advanced feature engineering from economic indicators (FRED) and NLP on FOMC statements.
Environmental & Spatial Analysis: Achieved 3rd Place in the Coral Reef Analysis challenge, applying advanced time-series and spatial analysis to forecast ecological trends.
Public Policy & Business Intelligence: Demonstrated rapid domain mastery by securing 3rd Place in the French Fiscal Challenge, analyzing 40 years of municipal tax data using DuckDB/SQL, despite no prior French language knowledge.

I am highly proficient in Machine Learning (Ensembles: XGBoost, LightGBM), Deep Learning (GNNs, BERT, FastText), Time Series Analysis, NLP, and Python.

Curiosity Meets Production-Ready Code

My approach is not confined to a single industry. I am driven by a deep curiosity and an ability to quickly assimilate new knowledge—evidenced by exploratory work in other fields like smart contract concepts (Solidity) and game development (GDevelop). I excel at translating sophisticated technical findings (using tools like SHAP for explainability) into clear, client-ready narratives for non-technical stakeholders.

I am eager to apply my versatile skills in environments that value both technical depth and strategic impact, particularly in Fintech, Applied AI Research, or high-stakes consulting engagements.

Core Expertise & Proven Achievements

My expertise spans the full lifecycle of Machine Learning and Advanced Analytics, applied across diverse, real-world domains:

Quantitative Finance & Time Series: Secured 2nd Place in the highly competitive FOMC Rate Prediction challenge by developing a robust multi-class ensemble model, integrating advanced feature engineering from economic indicators (FRED) and NLP on FOMC statements.
Environmental & Spatial Analysis: Achieved 3rd Place in the Coral Reef Analysis challenge, applying advanced time-series and spatial analysis to forecast ecological trends.
Public Policy & Business Intelligence: Demonstrated rapid domain mastery by securing 3rd Place in the French Fiscal Challenge, analyzing 40 years of municipal tax data using DuckDB/SQL, despite no prior French language knowledge.

I am highly proficient in Machine Learning (Ensembles: XGBoost, LightGBM), Deep Learning (GNNs, BERT, FastText), Time Series Analysis, NLP, and Python.