Augurisk — Climate & Societal Risk Platform
Platform that helps individuals and businesses assess the climate and societal risks associated with their properties. I built environmental and societal risk models covering floods, hurricanes, wildfires, earthquakes, air pollution, crime, socioeconomic risk, and health infrastructure — processing large-scale geospatial data and deploying models on cloud infrastructure to deliver property-level risk scores at scale.
Project Highlights
Built classification and regression models for 10+ environmental and societal risk categories.
Generated vector tilesets from large GeoJSON collections using tippecanoe, OpenLayers, and GDAL.
Deployed scientific models on AWS clusters (EC2 / EMR) for big-data processing.
Integrated third-party data sources: US Census Bureau, ACS, NIH, USGS, CODE.
Built automated anomaly detection with continuous performance tracking.
Explore More Work
Deep dive into other high-performance solutions.
Production AI system for call centers: an LLM-powered chatbot that assists agents with FAQ retrieval, intent classification, and response generation — paired with a full automatic speech recognition (ASR) pipeline that converts live voice calls into text. Continuously improved ML prediction and recommendation models in production, achieving a +20% performance gain through fine-tuning, cross-validation, and monitoring.
Platform that helps individuals and businesses assess the climate and societal risks associated with their properties. I built environmental and societal risk models covering floods, hurricanes, wildfires, earthquakes, air pollution, crime, socioeconomic risk, and health infrastructure — processing large-scale geospatial data and deploying models on cloud infrastructure to deliver property-level risk scores at scale.
Designed and automated a 0-cost AI-powered short-form content generation workflow for social media using n8n, Gemini, and Kokoro — enabling fully automated creation and orchestration of video scripts and content pipelines.
Conversational agent built over Augurisk's internal risk methodologies and scientific documentation. Implemented retrieval-augmented generation (RAG) using LangChain, FAISS vector search, and OpenAI to let internal teams ask natural-language questions about risk models and get grounded, citation-backed answers.
AI travel assistant generating personalized itineraries and recommendations for travelers in Morocco, with a scalable vision to cover all Moroccan cities. Combines LLM-driven planning with curated local knowledge to produce day-by-day trip plans tailored to user preferences.
Graph + ML fraud detection system identifying suspicious transaction patterns. Used Neo4j to model entity relationships across accounts and transactions, and scikit-learn classifiers (combining graph features with transactional features) to flag fraud rings invisible to row-by-row analysis.
End-to-end ML pipeline forecasting power consumption across 3 zones of Tetouan, Morocco — covering preprocessing, feature engineering, MLflow experiment tracking, containerization, and AWS deployment. Predictions are served through a Flask REST API.
Real-time financial data platform enabling live market ingestion and near real-time analytics for trading insights and decision-making. Built on Kafka for streaming and AWS (S3 + Glue + Athena on EC2) for the serverless analytics layer.
NLP analytics system tracking and visualizing public sentiment on Brexit over time using large-scale Twitter data and machine learning models. Produced descriptive statistics, time-series sentiment trends, and word clouds to surface dominant narratives.