About Me
Hi, I’m Husayn. I’m a Senior Data Scientist and Georgia Tech Ph.D. with more than ten years of experience turning messy data into actionable systems through predictive modeling, generative AI, and geospatial analytics.
I’ve built models, data pipelines, AI applications, and analytical tools that serve fields including environmental science, education, public health, engineering, and business operations. My projects range from satellite-based water-quality monitoring and climate analytics to computer vision and agentic RAG systems.
My foundation in civil engineering and hydrology taught me to look beyond isolated datasets and approach problems as interconnected systems. I bring this holistic perspective to building intelligent tools that can positively impact people, communities, the environment, and the way important decisions are made.
If you’re working on a problem where machine learning, AI, or geospatial analytics can make a meaningful difference, I’d be glad to connect.
Skills
Certifications
Portfolio
Remote Sensing & ML for Water Quality Assessment at Lake Lanier
Satellite-based ML workflow to monitor chlorophyll-a and assess water quality compliance for Lake Lanier.
Trial Conversion Prediction & MLOps
Turns early trial engagement into outreach priorities with XGBoost, automated AWS deployment, and data drift monitoring.
Autodesk Agentic RAG Chatbot
Evidence-grounded assistant for Autodesk product and workflow questions using hybrid retrieval, web verification, reranking, and evaluation.
Cobb County Building & Fire Code Agentic RAG Assistant
RAG-powered assistant for querying and retrieving information from Cobb County building and fire codes.
Telco Customer Churn Prediction & Retention Targeting System
Production-oriented churn prediction workflow with XGBoost, FastAPI, Streamlit, MLflow, and retention-value targeting.
AI-Automated Lead Generation Pipeline for Engineering Consultancy (RAG + Web Data)
RAG-powered pipeline to mine regulatory and web data, rank prospects, and surface qualified engineering leads.
Teacher Support Studio: Student Proficiency Prediction & AI Teacher Assistant
Predicts next-practice success from longitudinal learning data and turns explainable estimates into grounded, teacher-facing guidance.
AI-Powered Retinal Disease Detection with Deep Learning & Computer Vision Pipeline
TensorFlow-based CNN that classifies retinal images into disease categories to support early detection.
Multi-Agent LLM System for Automated Scientific Literature Review
Full-stack research automation app that searches Scopus, verifies DOIs, saves papers to Zotero, and exports cited review documents.
Voice-AI Customer Service Agent for Medical Lab Appointments
Voice AI system to automate medical lab appointment scheduling, rescheduling, and support calls.
Unsupervised Risk Modeling with K-Means on AMI Smart Meter Data
Segmented smart meters using k-means clustering on high-frequency AMI power-quality data to identify behavioral patterns linked to grid reliability and outage risk.
Public Health A/B Testing & Predictive Analytics (XGBoost)
Application of A/B testing and machine learning (XGBoost) to measure message impact and target individuals most likely to schedule appointments, with SHAP-based explainability.
Climate-Informed Crop Yield & Irrigation Projections for the ACF River Basin
Modeled climate-adjusted crop yields and irrigation demand to support long-term agricultural and water planning.
Continental-Scale Climate Bias-Correction, Projection, & Interactive Dashboard
Built bias-corrected climate projections and an interactive dashboard to explore multi-model climate futures.
SOCO 48-Hour Energy Demand Forecasting
End-to-end ML pipeline for forecasting hourly electricity demand using 10+ years of grid and weather data, achieving high-accuracy 48-hour load predictions.
Professional Experience
Senior Data Scientist (Part-Time)
- Developed an XGBoost lead-scoring system using weather, permit, valuation, and geospatial data, projecting a 12% revenue lift (~$450K per 600 leads) and improving precision-recall performance by 3.3x over a heuristic baseline. Added Platt calibration and SHAP for explainable per-lead insights.
- Built an explainable RAG platform to match project requirements with qualified engineering firms, indexing 7K+ public profiles into ~70K retrieval documents using PostgreSQL/pgvector, embeddings, hybrid search, reciprocal rank fusion, and cross-encoder reranking.
- Built an AI-powered lead-enrichment pipeline with Python, Google Gemini, and the Google Places API to surface high-value industrial prospects for an engineering consultancy, replacing manual prospecting with a scored outreach list and accelerating lead qualification.
Senior Data Scientist (Research Engineer)
- Developed bias-correction and spatiotemporal downscaling algorithms for 16 climate models across two emissions scenarios, producing 4-km daily climate projections while preserving temperature-precipitation correlations. Integrated climate, soil, and crop data into ML models to predict crop yield and irrigation demand and quantify long-range agricultural risks across the ACF River Basin.
- Ported and modernized the SAC-SMA hydrologic model from R to Python using OpenAI Codex, adding flexible time steps, units, and calibration objectives. Achieved 100% model fidelity through rigorous testing, enabling high-performance computational workflows to assess projected changes in Georgia's water availability through the year 2100.
- Developed and evaluated an artificial neural network (ANN) to predict daily solar radiation from meteorological data, reducing RMSE by up to 9.7% versus a calibrated empirical model and supporting long-term agricultural forecasting.
- Integrated and evaluated multi-sensor satellite and gridded climate datasets from NASA, JAXA, and NOAA, performing statistical error, bias, and spatial correlation analyses to assess data reliability for predictive modeling across the continental U.S.
- Containerized geospatial ML workflows using Docker, enabling reproducible deployment and execution of environmental monitoring and predictive modeling pipelines.
- Oversaw $400K+ in annual federal and state research funding as Assistant Director of the Georgia Water Resources Institute.
- Mentored graduate researchers in data science, programming, and ML methods, supporting publications and conference presentations.
Engineering Consultant (Part-Time)
- Built a TensorFlow-based decision algorithm combining neural networks and dynamic programming to select optimal membrane cleaning actions at 10-minute intervals, evaluating 74K+ system states to balance operational costs, equipment wear, and membrane performance.
- Developed and calibrated a data-driven decision support system for wastewater treatment operations, applying sensor time-series analysis and rule-based algorithms to automate membrane cleaning, detect anomalies, and handle edge cases across 32 operating scenarios.
- Supported AWS IoT integration for near-real-time monitoring of 24 operational KPIs at 60-second intervals, enabling data-driven insights into wastewater treatment performance and supporting the development of AI-powered operational systems.
- Developed physics-based and 3D modeling workflows and trained engineers in their adoption, improving design collaboration, reducing physical prototyping iterations, and shortening project delivery time by 20%.
Data Scientist (Graduate Research Assistant)
- Built and validated ML models to reconstruct missing environmental data, benchmarking 71 methods across 33 years of daily observations from 15 weather stations to improve data accuracy and statistical consistency for hydrologic and climate analyses.
Education
Ph.D., Civil Engineering
Master of Science, Civil Engineering
Bachelor of Science, Civil Engineering
Contact
Interested in connecting or discussing an opportunity? Send me a message and I’ll get back to you.