i'm a software engineer based in new york city working at meta.

work
cinder @ meta, software engineer
august 2023 — march 29, 2024 10:25:43 am
  • developing meta’s performance-oriented fork of python that includes several performance optimizations, including bytecode inline caching, eager evaluation of coroutines, a method-at-a-time jit, and an experimental byte code compiler that uses type-annotations to emit more efficient bytecode
ligen @ meta, software engineer
september 2021 — august 2023
  • developed a compiler that is written in rust and generates hyper-efficient direct- bindings between language pairs (c++, rust, python, haskell, go, java) that is used to improve performance of core parts of instagram and whatsapp tech stack by relying on lower-level libraries.
  • migrated several critical, highly-performance sensitive codepaths for instagram to use bindings generated by the compiler — including its database access layers, content distribution networks, and caching layers — totaling >300 million requests per second and responsible for >15% of instagram’s global cpu usage
  • worked on the entire stack of the compiler: frontend parsing, intermediate representation transformations, individual language code generation
  • designed and architected the implementation of poking and peeking complex data structures across the language boundary, including hash maps, hash sets, vectors, etc.
  • worked at an extremely low-level, looking at assembly generated by clang and performing unsafe memory modifications of data structures to improve performance
  • experienced in profiling c++ code
  • compiler output battle-tested with hundreds of millions of requests per second, performing equally or better than similar language standards (cxx.rs, cython, etc.) and powers every database query and media request on instagram
translation @ facebook, software engineer intern
may 2020 — july 2020
  • productionalized mbart multilingual neural machine translation model for usage within facebook’s stack of translation services
  • fine-tuned models to achieve significant increase in bleu scores
  • model significantly improved translation performance for underserved languages
hr analytics @ hca, software engineer intern
may 2019 — december 2019
  • designed a web application using react, golang, and microsoft sql server
  • application analyzed and predicted nurse turnover, transfers, and hiring using statistical models of an inverted labor market
  • application deployed to 180+ hospitals across the usa to manage 80k+ employees and continues to be used by regional and corporate HR VPs to facilitate overall company growth

© aditya pillai 2024

made with in nashville