Alex Anthraper - AI & Analytics Architect

Summary

AI & Analytics Architect with over 8 years of experience designing and deploying end-to-end Agentic AI, Generative AI, and Optimization solutions. Expert in architecting scalable cloud-native platforms on Azure and Databricks, integrating LLMs with enterprise systems to drive autonomous decision-making. Proven track record of technical ownership—from road-mapping to production deployment—delivering measurable value (over $60M impact) through robust MLOps, containerization, and cross-functional leadership.

Core Competencies

• Agentic Workflows & LLMs

• Generative AI Solutions

• Agentic BI & Automation

• Optimization Solvers

• Forecasting & Planning

• Resource Optimization

• Cloud Architecture (Azure)

• Databricks & Delta Lake

• Containerization (Docker/K8s)

• ML-Ops & API Development

• Data Observability & Governance

• Open-Source BI Stack

• Product Architecture

• Strategy & Road-mapping

• Team Leadership

• Stakeholder Management

• Cross-Functional Collab

• Value Realization

Professional Experience

Data Science Manager

Oct 2021 - Present

Americana Restaurants

UAE

Leading a team of 6 to architect and build autonomous AI agents, scalable ML platforms, and advanced analytics solutions for 10+ QSR brands.

Agentic AI & Automation

Conversational BI Platform: Architected a tool-using conversational agent (Python, FastAPI, LLMs) for Apache Superset. Enabled natural language queries of complex BI dashboards, democratizing data access.
Automated Store Segmentation Agent: Designed and deployed an autonomous agent for dynamic store segmentation. The agent plans and executes clustering tasks based on real-time price elasticity data.

Cloud Architecture & Infrastructure (Azure/Databricks)

Enterprise Delta Lake Backbone: Architected the core MLOps foundation. Implemented ETL pipelines with strict cataloging, profiling, and metadata definitions (Azure Data Factory, Databricks).
Open-Source BI Stack (AKS): Engineered a unified self-service analytics platform via containerized deployment on Kubernetes, integrating Spark and Superset to reduce licensing costs by 40%.
Data Observability Platform: Built a governance platform that auto-catalogs and profiles data quality, reducing discovery SLAs from weeks to minutes (FastAPI, Apache Drill, Svelte).

Forecasting & Analytics

Multivariate Revenue Forecasting: Achieved 90%+ accuracy across 2300 outlets using Time series/ML models (SARIMAX, XGBoost), improving supply chain planning.
Construction Timeline Prediction (Causal Modeling): Improved opening-date forecasts for 50+ projects using Causal inference and regression (Statsmodels), optimizing resource allocation.
Promo ROI Analysis: Optimized marketing spend by identifying top-performing campaigns using Causal inference and decision trees (Scikit-learn, PuLP).
Commodity Risk Management: Stabilized costs for a $1.5B+ portfolio using Binomial options pricing models and financial hedging strategies.

Optimization & Decision Science (Value Realization)

Differential Pricing Engine: Delivered $55M in revenue increase by engineering a price optimization system (Simulated Annealing) deployed on a FastAPI/Svelte stack on Databricks + Kubernetes.
Ant Brain - Staff Transport Optimization: Reduced transport costs by 14.2% using Learning Ant Colony Optimization (ACO). Selected as Americana's entry for the INFORMS 2025 Daniel H. Wagner Prize.
HD Driver Schedule Optimization: Reduced delivery costs by 11% across 1,800+ sites using a Mixed Integer Programming (MIP) model with quadratic constraints (PuLP, Gurobi).
Smart Vendor Selection: Achieved $9M annual savings by optimizing procurement via a MIP model with complex constraints for MOQ, capacity, and shipping costs.

Data Scientist

Jul 2019 - Oct 2021

Triton Middle East LLC

UAE

Customer Follow-Up Optimization: Boosted customer reach by 94% through automated CRM integration and intelligent agent-like scheduling systems.
Territory Design (Geospatial Clustering): Increased average daily customer visits per rep by 20% using Geospatial clustering (K-Means).
Delivery Routing Optimization: Reduced order-fulfilment SLA by 50% using Zone-based K-Means clustering + MILP optimization.
Credit Risk Prediction: Reduced unsecured credit exposure by 44% using an MLP neural network classifier.
Cross-Selling Recommendation Model: Increased average order value (AOV) by 8% through collaborative filtering recommendations.

Developer, Data Analyst

Aug 2015 - Jul 2019

Triton Middle East

UAE

Contributed to core systems development, saving 14,000+ man-hours through CRM workflow automation and streamlining analytics via ERP integration.

Education & Certifications

MSc, Applied Statistics in Finance

University of Strathclyde, UK Jul 2026 (expected)

CQF (Certificate in Quantitative Finance)

CQF Institute, UK Dec 2025 (expected)

Post Graduate Diploma in AI & ML

IIIT, India Nov 2021

Bachelor Program in Engineering

NIET, India Sep 2009

Leadership & Training

Power BI Trainer, Institute of Chartered Accountants of India (ICAI)
Fellow, Royal Statistical Society (UK)

Recognition & Publications

Award: Americana's Official Entry - 2025 INFORMS Daniel H. Wagner Prize.

White Papers: Differential Pricing Optimization, HD Driver Schedule Optimization, Employee Transportation Optimization

Tools & Technology

Python • Azure • Databricks • Kubernetes • Docker • PySpark/Spark • FastAPI • Apache Superset • Gurobi • PuLP • NetworkX • Scikit-learn • XGBoost • TensorFlow • Apache Trino • Apache Iceberg • Svelte • HTMX • SQL • N8N