Patrick Kelly

Data Platform Engineer · Orange County, CA
Data infrastructure, analytics engineering, and AI-ready data systems

[email protected] · linkedin.com/in/pjkelly82 · +1 562-505-0434

Data platform engineer with 20 years of experience building production systems across music, media, and technology. Focused on the full data stack — from ingestion architecture and warehouse design to semantic layers and AI-ready data products. Brings deep technical ownership and the organizational fluency to work effectively across engineering and business.

Experience

Staff Data Engineer · Rostrum Pacific

February 2025 – Present · Remote

Sets technical direction and architecture for the data engineering team at a music distribution and publishing company. Responsible for the full data platform across consumption, payments, and playlisting domains, from raw ingestion through to analytics, AI, and BI products.

• Designed and implemented a multi-source ingestion framework supporting 130+ data sources; architected the warehouse and data lake to serve analytics, AI, and business intelligence across the organization.
• Designed and built the first version of the company’s Analytics AI agent, including the semantic layer underpinning it: business-defined metrics, routing logic, and hybrid search, to ensure AI outputs are grounded in reliable, well-modeled data. Continues to collaborate with the AI engineering team on data-adjacent concerns.
• Partnered with executive leadership to deliver a company-wide BI solution, driving data availability, structure, and quality for organization-wide reporting.
• Applied a deliberate build-vs-buy approach to infrastructure decisions, replacing legacy tooling with purpose-built solutions where appropriate and negotiating vendor contracts and cloud capacity agreements to manage costs and reduce operational risk.
• Built a payment data pipeline standardizing DSP payment data across all sources, supporting accounting workflows and generating downloadable customer statements.
• Owned all infrastructure-as-code across data and application infrastructure, ensuring consistent, version-controlled environment management.
• Reduced Spotify consumption data latency from 48+ hours to under 24 hours through targeted pipeline optimization.
• Established proactive data health monitoring covering platform delivery status and data quality across all sources.
• Designed a data crawler framework for automated collection of data unavailable via standard delivery or API.

Data Engineering & AI Consultant · AnswerQuest

January 2025 – April 2025 · Remote (Contract)

AnswerQuest is an AI-powered communication platform for K–12 schools, using school-verified documents.

• Built the initial data pipeline infrastructure supporting ingestion and processing of school knowledge sources.
• Developed the AI chat back-end using Python, FastAPI, AWS Bedrock, LangChain, and LangGraph.

Lead Data Engineer, Technical Lead · Crush & Lovely

March 2006 – August 2024 · Remote

Data and platform lead at a digital agency serving enterprise clients across media, entertainment, and technology. Owned data architecture across a portfolio of concurrent client engagements for 18 years, spanning warehousing, pipeline infrastructure, data modeling, and analytics.

• Owned data warehouse architecture and modeling across client platforms, designing for analytical and operational workloads across relational and NoSQL databases.
• Built and operated ELT pipeline infrastructure using Dagster, dbt, and custom Python tooling; developed proprietary data connectors and extraction systems for sources without standard integration paths.
• Led data analysis engagements, working directly with client datasets to surface insights and identify optimization opportunities.
• Set technical direction across all engineering disciplines; served as the primary client-facing architecture and data lead, translating business requirements into platform design.
• Managed and mentored engineering teams across data, back-end, and DevOps disciplines across a multi-client portfolio.

Notable clients: Disney · ABC · NBC Universal · Comcast · IBM · American Heart Association · National Association of Broadcasters

Skills

Data Engineering Dagster · dbt · PySpark · AWS Glue · Pandas · Python

AI & Analytics Semantic layer design · LLM integration · hybrid search · named entity recognition · AWS Bedrock · LangChain · LangGraph

Cloud & Infrastructure AWS (Redshift, S3, Glue, Athena, Bedrock, Lambda, ECS, Kinesis, CloudWatch) · Terraform · GitHub Actions

Education

Applied Data Science Program Massachusetts Institute of Technology Professional Education · 2023

Leveraging AI for Effective Decision-Making: exploratory data analysis, data visualization, clustering, predictive analytics, decision trees, deep learning, and recommender systems.

Bachelor of Arts, Double Bass Performance Eastman School of Music, University of Rochester · 2000–2004