Akshansh Baliyan

About Akshansh Baliyan

Data Scientist & Business Intelligence Engineer at AWS with 4+ years of experience applying optimization, machine learning, and large-scale analytics across tech, financial services, and pharma.Currently, I develop constraint-based optimization models that power GenAI/ML infrastructure planning at AWS—balancing manufacturing capacity, power, space, and throughput across 200+ global data center sites and 52-week planning horizons.My work involves:• Multi-objective optimization across allocation combinations• Forecasting models achieving >97% accuracy (99.78% near-term)• Large-scale data pipelines processing 10M+ rows weekly• Recursive allocation algorithms with hierarchical constraint propagationPreviously at American Express, I built predictive models (Random Forest) to identify high-value customers and developed scalable insights packages using PySpark and Hive for 170+ global banks. At ZS, I designed cloud-native ETL pipelines (S3, Glue, Redshift) supporting $2B+ monthly pharma revenue.Impact highlights: 94.78% global utilization,$10M+ annual savings, 40+ hours/week automated.Education: MS in Management Information Systems & Analytics, University at Buffalo (3.93 GPA)Stack: Python, SQL, PySpark, Redshift, Airflow, Tableau, AWSAlways eager to tackle complex, ambiguous problems and build solutions that drive measurable impact. Open to connecting with curious minds and exploring new ideas.HOW TO CONTACT ME:Email: a••••••••@gmail.com

Experience

1. Business Intelligence Engineer

   Amazon Web Services (AWS)

   Sep 2024 — Present · Seattle, WA, US

   Led the development and deployment of data science solutions to optimize Amazon\'s global GenAI/ML infrastructure deployment, managing rack allocation and data center supportability across 150+ sites worldwide.Key Achievements:• Architected and implemented an end-to-end rack allocation system achieving 97% forecast accuracy and 94.78% global build utilization, driving $10M+ annual cost savings through improved infrastructure placement and manufacturing optimization• Designed and deployed a sophisticated Manufacturing Allocation Algorithm managing weekly capacity across 10+ products and 50+ clusters, incorporating 52-week forecasting and complex constraint optimization• Built comprehensive DC Throughput Allocation system integrating power, space, network, and manufacturing constraints to maximize data center supportability across global infrastructure• Developed interactive visualization dashboard enabling real-time simulation and audit capabilities for stakeholders, reducing manual planning effort by 40+ hours weeklyTechnical Implementation:• Engineered robust data pipeline processing 1M+ weekly records from 10+ internal sources using Python, SQL, AWS Redshift, S3, Airflow, and ECS• Implemented complex allocation logic handling combinations across multiple hierarchies (Cluster → AZ → Site) with dynamic prioritization• Created 2500+ lines of Python code and 50+ Redshift SQL scripts, orchestrated through enterprise toolsCross-functional Impact:• Enabled EC2 Capacity team\'s ML infrastructure allocation through automated supply data integration• Facilitated Finance team\'s revenue projections through precise infrastructure landing timeline visibility• Automated critical reporting for Network Capacity, BBU inventory planning, and Supply Chain operations

Education

• The Cambridge Schools

  Physics Chemistry Maths

  2002 — 2016

• SRM IST Chennai

  Bachelor of Technology (B.Tech.), Electronics and Communications Engineering Technology

• University at Buffalo School of Management, The State University of New York

  Master's degree, Management Information Systems - Analytics