HELLO THERE,

Cloud Infrastructure Automation

• Engineered cloud-native deployment system for academic analytics tools:
• Automated provisioning of 50+ R applications using ShinyProxy/Terraform/Docker
• Implemented LDAP/SSO integration with granular RBAC policies
• Reduced VM provisioning time by 65% through infrastructure-as-code practices

Learning Analytics Pipeline Development

• Architected real-time clickstream processing system handling 20M+ daily events:
• GCP dataflow pipeline with PubSub/BigQuery integration
• Apache Superset dashboards monitoring 150+ engagement metrics
• Anomaly detection model identifying classroom irregularities with 92% accuracy

Academic Research Modernization

• Transformed EDUCAUSE survey analysis through:
• Dimensional modeling on AWS Redshift with MVR refresh automation
• Natural language processing pipeline using SpaCy/NLTK
• LLM-powered Q&A system (Gemma2 9B/Ollama) with RAG architecture
• Developing cross-department Tableau dashboards serving 12 academic units

Cloud Data Fabric Development

• Collaborated with a Lead Data Architect to drive a 10 engineer team in building an AWS-based data fabric from scratch, accelerating the client’s time-to-market by 30% through:
• Custom ETL pipeline design with SCD Type-2 historization for accurate historical tracking
• Metadata-driven workflow management for dynamic ETL orchestration
• Real-time monitoring API using AWS Lambda & GraphQL for proactive failure detection

ML Engineering & Analytics

• Developed a scalable customer segmentation model using Spark ML/K-Means, leveraging PCA-driven feature engineering
• Implemented a forecasting engine on a GCP Data Lake comparing:
• ARIMA for capturing seasonal trends in time-series forecasting
• Random Forest & XGBoost for enhanced prediction accuracy
• Designed a phased rollout strategy for ML model deployment, ensuring gradual adaptation and A/B testing

Cloud Infrastructure Automation

• Reduced cloud provisioning time by 60% using Terraform-based Infrastructure as Code (IaC) templates for AWS/GCP
• Built a serverless cost optimization framework for dynamic EC2 resource allocation, leveraging:
• Spot Instance optimization for EMR clusters, reducing compute costs
• Auto-scaling strategies for dynamic resource allocation based on demand

Data Governance & Security Compliance

• Architected a metadata-driven Data Quality (DQ) framework integrating Great Expectations & Deequ for automated data validation
• Implemented a scalable PII protection system with:
• AWS Macie-powered sensitive data detection for compliance with GDPR/CCPA
• Format-preserving encryption (FPE-FF1) in Python/Scala to maintain referential integrity

ML Engineering

• Scaled TensorFlow training pipelines using Python/Ray across 6-node cluster, reducing computation time by 10%
• Developed production ML evaluation CLI with MLFlow/Python that automated feature scoring, reducing manual engineering effort by 25%
• Implemented market segmentation models (Decision Trees/Random Forest) for personalized product recommendations

Data Engineering

• Designed and developed a Linux-based command-line tool for automated data ingestion and storage:
• Ingested enterprise, asset, meteorological, and failure data from 70+ geo-distributed locations of a renewable energy giant
• Structured data into a multi-zoned AWS data lake, ensuring efficient organization and retrieval
• Implemented user-specified granularity, enabling data science teams to build historical datasets
• Powered KPI-driven dashboards for energy grid performance and downtime analytics
• Integrated Spark Streaming into ETL pipelines and migrated from Spark 1.4 to Spark 2.4:
• Processed ~100GB of IoT data daily from 70+ wind and solar farms
• Structured data in a multi-zoned AWS data lake for consistent availability
• Enhanced processing speed and scalability through Spark 2.4 migration
• Developed a command-line interface for CI/CD automation:
• Automated Docker image creation and registry push
• Integrated with a remote Jenkins server to trigger CI/CD build pipelines
• Standardized and automated containerized deployment workflows