Data Engineering and Architecture
I design and engineer scalable and reliable data platforms end to end. I have extensive experience and expertise with Databricks, Snowflake, AWS and Azure.
Data Architect & Engineering Lead
Designing and building data platforms, ML and AI systems that scale to billions of rows and turn data into decisions.

Built products trusted by Hulu/Disney, Roblox, Zoom, Klarna, Compass, Okta
For infrastructure cost visibility and optimization, across $1B+ in annualized cloud spend.
500%
Increase in scale of data processing pipelines to 2.5 billion rows per source, 30% faster and at 15% lower cost.
$600K saved
Built outlier detection to detect cost spikes, consuming billions of data points. Saved $600K for Zoom in one detected incident.
79% reduction in latency
Rearchitected the data processing system to achieve 79% lower latency and an 80% processing speed-up.
I design and engineer scalable and reliable data platforms end to end. I have extensive experience and expertise with Databricks, Snowflake, AWS and Azure.
I turn data into decisions. I have built time-series forecasting with high accuracy, anomaly detection, infrastructure optimization recommendations, and Generative AI (RAG and ReAct).
Architected the AWS cost pipelines that became the core of the platform, then extended them to Azure, Google Cloud, Kubernetes container costs and rule-based business hierarchies, at up to 2.5 billion rows per run. Added container rightsizing from actual versus allocated resources, and tuned Spark to cut runtime 30% and cost 15%.
Built and scaled anomaly detection over 6–8 billion rows of cost data, catching $600K of runaway spend for Zoom. Engineered an ensemble forecasting service that produces 24-month forecasts at every level of the business hierarchy, tuned with cross-validation and back-testing. A Generative AI assistant built on ReAct agents answers questions about financial data in plain language.
Scaled the data processing pipelines to 50% more processing capacity and over 1 billion rows per run. Led the architecture, benchmarks and proof of concept to replace MongoDB with a sharded ClickHouse warehouse: 79% lower median query latency, and data loads cut from five days to one. Run with Claude Code, Skills and hooks for reproducible, evidence-backed decisions.
Simulated AWS's Reserved Instance application algorithm on hourly EC2 inventory and optimized convertible RI exchanges with a mixed-integer linear program, plus near-real-time cost estimates that bridge the 2–3 day billing lag. A key capability in closing Intuit as a customer.
Led the architecture evaluation, cost and resourcing estimates, and a production pilot that ran one customer's cost pipeline end to end on Databricks.
Leading the ClickHouse analytics warehouse modernization, and directed the customer migration from the legacy platform with 100% retention.
Joined as a founding team member and grew into architect and lead of the data platform behind FinOps products used by Hulu/Disney, Intuit, Roblox, Okta and Zoom, leading a pod of up to five engineers.
Fulbright Scholar: graduate study in the United States under the Fulbright Program.
Built business intelligence: SSAS cubes and MDX, SSRS reporting tools, Master Data Services customization, and reporting queries tuned to respond in under a second.
iOS application and game development for iPhone and iPad; received the Rising Star award.
Data Processing: Apache Spark, PySpark, Scala, Apache Airflow, Kubeflow
AWS: Amazon EMR, Amazon S3, AWS Glue, Amazon Athena, Amazon Redshift, Amazon Kinesis, AWS Lambda, Amazon EKS
Databricks: Lakeflow Pipelines, Jobs, Unity Catalog, Declarative Automation Bundles, Delta Lake
Snowflake: Virtual Warehouses, Snowpipe, Streams and Tasks, Time Travel, Secure Data Sharing
Azure: Azure Data Lake Storage (ADLS), Azure Blob Storage, Azure Databricks, Azure Kubernetes Service (AKS), Microsoft Entra ID
Warehouses and Databases: ClickHouse, PostgreSQL, Amazon Redshift, Google BigQuery, DuckDB
Technical writing on data engineering, Spark, ClickHouse and AI tooling.
Email or LinkedIn is the fastest way to get in touch.