Data & Platform Engineering Leader

Designing and building offline-first mobile AI applications — local state persistence, pre-computed embeddings shipped on-device for sub-millisecond similarity search at zero API cost, and an async sync queue that defers LLM inference to when connectivity returns. Works in airplane mode, recovers cleanly on reconnect, and keeps user data local by default.

Building LLM features as dynamic context pipelines rather than static system prompts — system messages assembled at runtime from live application state with token-aware truncation, so responses stay grounded in current data rather than the state of the world when the prompt was written.

Multi-model LLM routing where cheap models handle the hot path and larger models escalate only when the signal justifies the spend — paired with deterministic fallbacks so the system keeps running on rule-based logic when the model call fails or times out, rather than hard-blocking the user.

Advising teams on data platform strategy and AI adoption — where to start, what to build first, and what is worth the investment. Patent filed in the on-device AI space.

Owned data strategy, data engineering, and the Cisco CX Agents deployed on customer devices across Cisco's product portfolio — the data layer powering Cisco's first cross-product customer experience cloud product. The product scaled to hundreds of millions in ARR, serving thousands of global enterprise customers.

Partnered with Enterprise IT on enterprise data and with engineering teams across the full product portfolio to bring in telemetry. Also delivered fully air-gapped and on-premise variants for government and regulated customers where cloud-connected telemetry was not permitted.

The team turned raw telemetry into curated insights that contributed to meaningful growth in annual services revenue and a significant reduction in renewals cycle time — though sales execution played a big role too.

Built the data team through direct hires, internal upskilling, and professional services partners. Led the multi-year legacy-to-cloud-native migration of the data layer platforms owned within Customer Experience Engineering — necessary work that touched every layer of the stack the team owned.

Built Dropbox's enterprise data organization under the CIO — from a small founding team to a mature, cross-functional group. The team covered finance, HR, marketing, sales, customer experience, and business operations data. Stood up the first enterprise analytics platform, and partnered with the product data org on the company's enterprise reporting capability.

Early Snowflake adopter — the team's work during Dropbox's formative years with Snowflake later led to a seat on Snowflake's Black Diamond Executive Council and a finalist spot at the inaugural Snowflake Data Drivers Awards in 2019.

The team's work on data strategy was central to Dropbox's public offering — a hyper-growth period when overall company revenue roughly doubled. Partnered with Finance, Growth, HR, and Business Strategy to certify tier-one data assets for external reporting — took months of getting every team that touched revenue data to agree on definitions, but without that work the filings would not have held up to scrutiny.

Ran data integrations across a broad SaaS footprint spanning finance, HR, sales, and marketing systems — one of the larger enterprise integration scopes at the company at that time. Not glamorous work, but the kind of thing that quietly makes the rest of the analytics platform possible.

First data engineering leadership hire. Rebuilt the function — consolidated BI and data engineering, set up governance, and handled the data integration work for an acquisition during a strong revenue growth period.

Delivered the first unified user analytics across the full product portfolio. Introduced self-service BI and the first executive dashboard.

Helped prepare for the company's first public analysts' day — the team certified every externally reported metric. The data credibility clearly played a part in what came next.

Managed data pipelines and production engineering for Yahoo's advertising and audience businesses — hyperscale systems processing billions of events a day across hundreds of petabytes, powering a multi-billion-dollar advertising business.

Built the first unified advertising data mart used by a large internal analyst and operator community. The underlying analytics solution was a Gartner BI Award finalist; team members won multiple U.S. patents for early Hadoop integration work.

Pushed for early enterprise adoption of open-source big-data stacks — Hadoop, Spark, Storm, Druid, Kafka — before most of them were considered production-ready. Also led a significant consolidation of legacy reporting platforms.

Infrastructure at scale: thousands of servers, hundreds of production database clusters, hundreds of petabytes of storage. Led complex multi-data-center migrations with strict security requirements.

Database administration and production operations across Oracle, MySQL, and SQL Server environments. Designed large-scale analytical databases, implemented ITIL practices, and ran enterprise BI environments.