Portfolio

Selected case studies across enterprise data, digital systems, and applied AI.

These case studies show systems thinking, measurable execution, and the ability to translate technical detail into business value.

Enterprise data systems

SLB Enterprise Data Engineering

Schlumberger (SLB) · August 2022 - May 2025

Scaled master-data and integration workflows across legacy systems to improve reporting trust, migration readiness, and analytics quality.

Business context

The operating environment spanned fragmented systems, inconsistent records, and migration pressure. The goal was to make enterprise data reliable enough for reporting, governance, and future analytics.

Approach

  • Optimized 40+ ETL/ELT workflows spanning 10+ source systems.
  • Built archival and monitoring patterns to improve troubleshooting visibility.
  • Introduced 25+ validation checks to catch quality issues before downstream consumption.
  • Added SOAP-based services for real-time material number reservation and synchronization.

Impact

  • 40M+ records transformed into unified golden records.
  • 20% increase in MDM system reliability.
  • 99% data readiness for downstream analytics and ML use cases.
  • 30% reduction in synchronization time for material number workflows.

Stack

Python SQL / PLSQL Informatica SAP Azure DevOps Data Governance
Digital operating model

TMHNA Digital Momentum

Kelley MSIS CORE Final Project · November 2025 - December 2025

Designed a target-state recommendation for TMHNA to align systems, processes, and finance data across TMH, Raymond, and THD.

Business context

A $7B enterprise needed a clearer target state for finance intelligence across multiple ERP and analytics platforms, without losing flexibility as reporting requirements matured.

Approach

  • Defined a target-state financial intelligence model spanning SAP S/4HANA, SAP ECC, and Snowflake.
  • Focused on KPI standardization, scalable data modeling, and analytics-ready datasets.
  • Balanced business storytelling with technical feasibility for executive and systems stakeholders.

Impact

  • 3rd Place finish in the Kelley MSIS CORE final project.
  • Created a clearer enterprise reporting blueprint for multi-system harmonization.
  • Positioned finance data for more reliable KPI definitions and future-scale analytics use.

Stack

SAP S/4HANA SAP ECC Snowflake Enterprise Architecture Business Case Design
Cloud transformation

Cloud Migration & Multi-Cloud Integration Strategy

Kelley MSIS Coursework · October 2025 - November 2025

Designed a 100% cloud migration roadmap and integration architecture for a mixed enterprise application estate spanning ERP, CRM, HCM, custom .NET, and IoT systems.

Business context

The brief required a practical migration path for a fragmented legacy landscape while balancing platform fit, real-time orchestration, global scalability, security, and cost discipline.

Approach

  • Mapped the estate across ERP, CRM, HCM, custom .NET applications, and IoT workloads.
  • Selected SaaS, PaaS, and IaaS target-state platforms across OCI, Azure, Salesforce, and Workday.
  • Designed an iPaaS-led multi-cloud integration layer for real-time orchestration and resiliency.
  • Built security, HA / DR, and cost-optimization controls into the proposed architecture.

Impact

  • Produced a full cloud migration plan for the target enterprise landscape.
  • Created a multi-cloud integration blueprint aligned to real-time data movement and global scale.
  • Connected platform selection with governance, resilience, and implementation priorities.

Stack

OCI Azure Salesforce Workday iPaaS Integration Architecture
Applied AI research

Pothole Detection Using Deep Learning

Springer Publication · October 2021 - February 2022

Built an ADAS-oriented pothole detection workflow focused on data curation, benchmarking, and practical model selection.

Business context

The project required a model that could balance accuracy and inference efficiency in a real-time road-safety context.

Approach

  • Built and augmented a dataset of 1,995 Indian road images.
  • Benchmarked YOLOv5 variants against Faster-RCNN with a ResNet101 backbone.
  • Selected YOLOv5m based on balanced precision-recall trade-offs and practical deployment speed.

Impact

  • 82% accuracy for the chosen model variant.
  • Produced a research-backed system design for ADAS-aligned pothole detection.
  • Demonstrated fluency in experiment design, evaluation, and model selection trade-offs.

Stack

Python YOLOv5 Faster-RCNN Computer Vision Model Evaluation