Core Platform Engineering: Helped scale Galileo’s infrastructure and engineering culture during a period of rapid growth.

* Architect and main developer behind Agent Control: https://agentcontrol.dev/

* Defined and enforced best practices for Python performance, helping drive down latencies and costs

* Built a client-side router for LLM inference, boosting utilization by 40% and cutting tail latency by 70%. https://galileo.ai/blog/how-we-boosted-gpu-utilization-by-40-with-redis-lua

* Created and open-sourced gcache, a product engineer friendly caching framework - https://github.com/rungalileo/gcache

* Built an internal load-testing framework to catch performance regressions

Helping Safety Systems team deliver safe AI products.

● ChatGPT: Developed model refusals and disclaimers around US elections and high-profile current events:

○ https://bit.ly/4chwHRH

○ https://openai.com/index/how-openai-is-approaching-2024-worldwide-elections

● Learned PySpark to create and optimize pipelines trawling through many terabytes of data, as well as for

some internal dataset analysis tools

● Optimized GPU-heavy jobs by 10x

● Helped with voice mode launch: https://bit.ly/4dv45WJ

Helped the DashPass team (~25 people) scale by spearheading the adoption of new technologies and best practices. Developed tech vision and led the transformation of a hard-to-maintain codebase into a modular, scalable platform. Helped other teams scale by standardizing and providing solutions for caching.

* Integrated Cadence to streamline partnership integrations, expediting the financial partnership process. https://doordash.engineering/2022/05/18/enabling-faster-financial-partnership-integrations-using-cadence/

* Adopted a functional testing framework to improve code coverage and software quality. https://doordash.engineering/2022/11/01/how-to-boost-code-coverage-with-functional-testing/

* Established and refined key engineering processes including on-call protocols, design reviews, coding standards, and architectural guidelines.

* Led a multi-quarter initiative to re-architect DashPass's core business logic into a plug-in-based platform, significantly enhancing engineering speed, code quality, and system reliability.

* Developed a robust, multi-layered caching library in Kotlin to optimize DashPass's backend performance, a solution later adopted across multiple teams and services within DoorDash. https://bit.ly/47SzdM9