Python Language Services and Experience Team* Starting the company’s investment in Python developer experience, creating team roadmapsSQL Native Code Generation Team* Designed protocol to compile generated code at runtime and dynamically load back to running process. The system allowed users to have up-to-date runtime dependencies while minimizing latency overheads.* Rearchitected codebase dependencies to suit multiple build systems, unblocked the team to open source along with our partner team’s data engine accelerator library.* Drove meetings across partner teams to convince organization leaders that code generation should be pursued in conjunction with interpretative data engine evaluation, resolved organizational misalignment on design priorities.* Set up open source continuous integration pipelines across multiple operating systems and compilertoolchains to maximize code portability.* Created a data serialization and deserialization library to help onboard Spark hash-joins to the new C++ backend, enabled the columnar-based data engine process to communicate with row-based data engines.Server LLVM Compilers Team* Brought the llvm lld linker to feature parity with the binutils gold linker, deployed lld at scale. The migration yielded 30 - 50% link time performance win.* Analyzed workflow performance and enable optimization techniques such as link time optimizations and profile guided optimizations, all the while balancing the tradeoffs between build time, runtime performance, and binary size. Hand tuned one of the largest C++ services and obtained around 10 million dollars in capacity savings.* Organized a cross-functional effort to prevent build problems due to static linking relocation overflows, leveraged the ELF format and added linker features to extend the possible code size. Examples include: improved cross-module LTO DWARF de-duplication, reordering symbols based on relocation distance, etc.

• Teach classes on low-level computer systems and assembly programming• Evaluate and provide insightful feedback on homework and programming projects for the 200-level course Computer Systems Fundamentals• Hold office hours to ensure students understand essential computer system concepts• Awarded the annual Outstanding Course Assistant Award in Spring 2017 by the Johns Hopkins University School of Engineering

• Deployed an automated solution to every machine in the Facebook network to better understand internal encrypted traffic by incorporating entropy metrics and information theory, enabling network security engineers and technical program managers to identify and prioritize possible security vulnerabilities• Consolidated multiple data sources and persisted data about plaintext traffic to track progress of migrating plaintext traffic to encrypted transport protocols using Presto and MySQL• Performed kernel programming to tag TCP retransmit packets and used eBPF to create traffic control filters

• Launched a full-stack solution to enable a user-friendly API for internal users to interact with backend servers via a Javascript front-end and a Java backend• Created new integrations with company ChatBot for streamlined database queries for the backend data management team• Managed transatlantic quality assurance development process

• Led a team of three in investigating the nature of word inflections in 800 different languages through machine learning techniques• Configured supervised learning algorithm and processed data for training purpose using Python and Shell Scripting