Jan 2020 – Present · 6 yrs 5 mos

• Proposed, designed and led the implementation of Flyte (flyte.org) at Lyft, which aims at democratizing large scale machine learning and data processing.
• Flyte is a multi-cloud, scalable, kubernetes native engine that makes it simple to build, operate and scale Data processing and Machine learning pipelines.
• Flyte is used at Lyft by 400+ DAU and is being adopted by multiple companies

Sep 2019 – Dec 2020 · 1 yr 3 mos

• Lead designer of next generation Machine Learning Platform at Lyft. Leading a cross organization effort that spans 30+ platform engineers.
• Leading the design of a large scale distributed training platform (using Flyte) for improving the velocity of engineers and researchers.

Feb 2017 – Dec 2019 · 2 yrs 10 mos

• Architect, Batch and Streaming pipelines on Kubernetes
• Built Flyte, which has helped deliver models and processes that impact Lyft’s revenue in millions of dollars. It powers critical backend data flow operations across all organizations.
• Proposed and led the implementation of Flink on Kubernetes at Lyft now OSS at flinkk8soperator and lead the implementation of Spark on Kubernetes.
• Drove many projects across 20+ engineers to improve reliability of data infrastructure at Lyft and Lyft L5.

Feb 2016 – Feb 2017 · 1 yr

• Lead a team to deliver a system capable of estimating traffic from Lyft driver cars, which are driving around in the country. This improves the ETA accuracy of the provided ETA for the customers, improving retention and also improving efficiency of the Lyft in general.
• Lead team to build an ensemble model to correct the bias within the model and personalize the models per driver, hour of the day etc.
• Designed and implemented a stream Map matching algorithm using Hidden Markov Models, that reduces the noise observed in GPS coordinates, which is the most important signal and significantly improves efficiency across Lyft.

Jan 2015 – Feb 2016 · 1 yr 1 mo · Greater Seattle Area

• Lead Engineer on Block Storage product for Oracle's new IaaS and BareMetal offering.
• Designed a scalable control plane and designed the unique virtualization scheme used for the data plane.
• Designed the interaction model and Microservices architecture for Block Storage control plane
• Created Anti-Entropy systems that would ensure that the intended state and the actual state do not differ considerably
• Implemented multiple modules and best practices that helped developing a V1 product at a brisk pace.
• Implemented extremely low footprint, efficient C++ service to handle control plane interaction with the data plane.
• Implemented a Workflow client library, which provided a simple programming model, abstracting concurrency, session/lease management from the customers of the library
• Designed and implemented a generic distributed Job Scheduling framework, to allow automation of fault-tolerant cyclic-operational jobs.
• Created a simple devops automation framework, that allowed all developers to be more productive. This abstracted the build system, code review process etc used within Oracle Public cloud. This was done as POC (in a short period) and showed the efficacy of having such a system. This triggered the organization to dedicate a team to build out the "Build/Code" management infrastructure using the same model as proposed.
• Designed and implemented a Local Per host Brownout prevention library, which is controlled by rules. The rules can be tweaked dynamically.
• https://cloud.oracle.com/bare-metal-storage

Jan 2014 – Jan 2015 · 1 yr

• Lead engineer leading the Backend systems for Amazon Maps. Functions include
• Designing the future version of Amazon Maps Data delivery platform
• Highly scalable delivery of rendered map data with industry leading customer experience
• Managing, ingesting and transforming of Geospatial Data for High performance Map Generation

Feb 2012 – Jan 2014 · 1 yr 11 mos

• Amazon Digital Products.
• Worked on the Map Control Developer API and other Geo-API's on Kindle Fire HD
• Worked on Amazon Maps.

Jul 2009 – Jan 2012 · 2 yrs 6 mos

• Transportation Route Optimization systems.
• Worked as a Lead engineer on
• Inbound Inventory movements
• Warehouse transfers
• Transportation document interchange systems
• High performance cost modeling for Transportation networks

Jun 2008 – Aug 2008 · 2 mos · Greater Chicago Area

• Financial Technology Analyst Program - Intern.
• Worked with the High Frequency Trading and Market making group at Citadel

Jul 2005 – Jul 2007 · 2 yrs

• Part of the team that built an Anti Money Laundering System for banks.
• Lead a team to create an automated behavioral detection platform for Suspicious activity detection