Vlad Tanov

About Vlad Tanov

Passionate about analyzing and solving complex, non-routine problems through innovative algorithms and creative solutions. I transform challenging analytical problems into actionable business outcomes that drive customer engagement and revenue growth at scale.8+ Years Experience in Python, Spark, SQL, and AWS cloud technologies as applied scientist/engineer/analyst • Full-Stack AI Development - End-to-end solution ownership including GraphRAG systems, Agentic AI WebSearch tools with specialist LLMs via supervised fine-tuning of Llama 3.2B foundational model, demonstrating complete AI application lifecycle management • Causal Inference Framework - Advanced four stage CausalForestDML framework with downstream integration via S3 and Redshift Spectrum • Anomaly Detection Systems - Deep Learning (LSTM) with contextualized web search using Brave API • MLDev/MLOps & Production Systems - SageMaker platforms deployed via CI/CD pipelines for AI/ML model training/inference • Optimization Software (Python), and migrating from Relational DW (AWS Redshift and RDS - SQL) to Data Lake (AWS EMR, Glue and S3 - PySpark and data lakes) • Big Data Pipelines - ETLs (SQL and/or Scopus Author ID Publications Data-Centric Optimization Approach for Small, Imbalanced Datasetshttps://jios.foi.hr/index.php/jios/article/view/1875• Machine Learning:http://imajor.info/LDA/Activities.html• A Nonsymmetric Nash-Riccati Equation And Decoupled Schemes for A Stabilizing Iteratively computation the Nash equilibrium points in the two-player positive Computing the Nash equilibrium for LQ games on positive systems

Experience

1. Applied Scientist Ii

   Amazon Web Services (AWS)

   Jun 2021 — Present

   Customer Journey Causal Inference Model - Developed breakthrough four-stage Single-Treatment DML pipeline for AWS\'s Customer Journey, enabling precise recommendations and resource allocation decisions. Successfully deployed production system serving 75K+ customers with personalized engagement strategies.2. 🤖 Full-Stack AI Development - Evolved from specialized data science to end-to-end solution ownership, building- GraphRAG systems - Agentic AI WebSearch tool - Supervised fine-tuning of Llama 3.2B foundational model, demonstrating complete AI application lifecycle management.3. Amazon Machine Learning Conference (AMLC) 2025 Workshop Leadership - Co-organized \"Foundational Models for Tabular Data\" workshop at AMLC, featuring keynote presentations and 12 innovative research projects, advancing Amazon\'s capabilities in structured data processing and LLM integration.2024:1. Usage Adoption System - Developed a customer\'s adoption system that measures services and their features variation in usage over time using metering data. Users are ranked and evaluated against each other based on historical metering usage. 🤖 Engineered a Sage Maker Machine Learning Dev platform, where Data/Applied Scientist could perform EDA, build and deploy models. The cloud infrastructure was developed as CI/CD, or infrastructure as code3. 🤖 Engineered Amazon MLOps - Automated Model Training and Inference Lambda Python Pipelines, including distributed machine learning

Education

• Sofia University St. Kliment Ohridski

  Doctoral Program (PhD), Data Science - Optimization

• George Mason University – Costello College of Business

  Bachelor’s Degree, Finance

• George Mason University - College of Engineering and Computing

  Master’s Degree, Data Analytics Engineering - Predictive Analytics

• Northern Virginia Community College

  Associate of Science, Business Administration