Building systems that endure.
I'm Madhan Selvam — a principal engineer who architects the data platforms enterprises run on. For 20+ years I've taken mission-critical systems from Mainframe to Hadoop to governed, AI-ready Lakehouses on AWS and Azure Databricks, across consumer goods, retail, finance and healthcare.
- PepsiCoArchitected A&M Hub — 50+ sources unified into governed Gold data products across 20 markets — and the data foundation for Integrated Business Planning across PBNA, PBUS and CAN.
- State StreetMoved a regulated AML and sanctions platform from on-prem Hadoop to an AWS Databricks Lakehouse: 70% faster, 40% lower storage cost.
- NikeBuilt and led member and commerce platforms — 10+ PB from 30+ sources serving 350M+ members, with $4M+ in cloud savings.
- Banking & healthcareA decade of Mainframe and Informatica integration at TCS, Cognizant and Aroghia for Citi, Bank of America, Union Bank, Anthem and Sutter Health.
Architecture principles
- Own the target architecture end to end — Lakehouse, streaming and batch platforms designed for scale, cost and AI/ML readiness, not just the next release.
- Set the technical direction: architectural guardrails, reference patterns and enterprise data models that teams across the organization adopt and build on.
- Lead from the code: hands-on in design reviews, critical-path implementation and performance tuning, raising the engineering bar through mentorship.
- Turn business strategy into platform decisions — aligning executives, product, data science and engineering on what to build and why.
- De-risk modernization with evidence: technical due diligence, POCs and phased roadmaps that move production platforms forward without disruption.
- Govern by design — lineage, quality, observability and privacy (GDPR, CCPA, DPA, HIPAA) engineered into the platform from day one.
Beyond work
After hours I run a personal R&D lab on where data engineering is heading — agents, copilots and open Lakehouse patterns — and publish what I learn as open blueprints.
ls ~/ai-lab
- claude-data-engineering-skills/Claude skills for ETL, schema inference, SQL generation and pipeline debugging
- data-engineering-copilot/AI-assisted workspace for PySpark, Airflow, dbt and data quality
- modern-lakehouse-patterns/Medallion, Delta and Iceberg, streaming and governance patterns
- spark-performance-lab/Query plans, AQE, Delta tuning and cluster-sizing benchmarks
- data-platforms-blueprints/Terraform, Kubernetes, CI/CD and observability for data platforms
cat learning.log
Harvard Business School leadership & strategy (2022–2024) · Databricks Platform Architect, GenAI and Azure (2026)
Bachelor's Degree in Engineering
Information Technology
Manonmaniam Sundaranar University
Tirunelveli, Tamil Nadu, India · 2004
Harvard Business School
Certificates · 2022–2024