• Architected ensemble forecasting system - Reduced pipeline runtime 55% (20h → 9h) while supporting 6+ concurrent model variants processing 100K+ weekly SKU forecasts.

• Migrated critical ML infrastructure to dedicated AWS segmented accounts using Terraform IaC, improving system isolation by 40% and maintaining 99%+ uptime across dev/staging/production environments.

• Optimized cloud resource utilization by 30% - Implemented shared preprocessing architecture reducing DPU consumption by 30%, achieving $50K+ annual cost savings.

• Built scalable ETL pipelines processing terabytes of Snowflake data to S3 using PySpark and AWS Glue, creating unified feature stores for time-series forecasting models.

• Designed event-driven workflows with SNS/SQS messaging, replacing monolithic structures with modular, fault-tolerant architecture and implementing automated data quality monitoring.

• Collaborated with data science teams to productionize research models (NPTS, MQCNN, WaveNet), translating experimental code into production ML pipelines with comprehensive documentation.

• Implemented comprehensive data quality monitoring with automated alerting for forecast anomalies, SKU drops >1%, and week-over-week variance exceeding 10% thresholds.

• Implemented S3 lifecycle management and Glacier archiving policies, automatically transitioning data >90 days old and reducing storage costs by 60-80%.

• Optimized large-scale data processing using AWS Glue and PySpark, implementing broadcast join optimizations and dynamic resource scaling to eliminate memory errors in multi-GB dataset operations.

Customer Reservation System / 2020 - 2022

• Designed and developed software to automate the scheduling of customer orders, significantly reducing order dwell while maintaining an on-time delivery rate of >90%.

• Managed containerized applications using Docker, Kubernetes & Helm in a CI / CD deployment process.

• Operated in a Kanban environment investigating, tech spec’ing and implementing solutions to allow our supply chain operations to expand internationally.

• Leveraged event-driven architecture (Kafka) to build cross-functional features that enhanced scheduling automation.

• Designed and built REST APIs for order and delivery scheduling.

‘Waygrade’ (Warehouse Returns Grading System) / 2019 - 2020

• Developed an in-house, warehouse facing application using React and Redux.

• Built a REST API for that warehouse facing application using Symphony and PHP.

• Enhanced the deploy process of the team by separating the front end and API from a single monolith repository into a decoupled application.

• Assessed usage and performance of the application in warehouse and assembled feedback into actionable information to improve the application.

• Integrated AWS S3 and Elastic Transcoder to convert video files into multiple, streamable formats

• Implemented full authorization and authentication using Auth0 to secure both a client side application and API

• Built a normalized database designed to house a diverse range of multiple media content

• Designed and built an in-house facing admin panel

• Oversaw student, staff & faculty access to campus buildings

• Executed and coordinated badging events for Berklee

• Managed student records resulting in maximized office efficiency

• Consulted students on how to resolve BPS involved issues