Vivek Nayak
Machine Learning Engineer at Meta
California, United States
Summary
👤
Senior🎓
Top SchoolVivek Nayak is a machine learning engineer with nine years of experience specializing in efficient LLM inference and on-device optimization. Currently at Meta, he focuses on accelerating LLMs for edge devices through ML compiler and runtime innovations, building on prior roles where he productionized TensorRT-LLM + Triton stacks and implemented quantization and KV cache reuse for Llama-3 models. At Capital One he delivered low-level fused kernels and trained SOTA adapters like Medusa-1 and Eagle, combining systems engineering with model distillation techniques. His research background at NYU included sequence-parallel QLoRA for long-context fine-tuning, reflecting a strong blend of research and production expertise. Comfortable across compilers, runtimes, and training kernels, he has moved models from prototype to highly optimized inference on constrained hardware. Per Aspera Ad Astra—he brings a pragmatic, performance-first approach to squeezing real-world efficiency from large models.
9 years of coding experience
4 years of employment as a software developer
BITS Pilani, Birla Institute of Technology and ScienceMaster of Science - MS Computer Science, Master of Science - MS Computer Science at New York University
Hindi, English
Github Skills (13)
kindle6
python6
mobi5
calibre5
transformers5
machine-learning4
deep-learning4
parallel4
pytorch4
huggingface-transformers3
language-model3
gpt2
deepspeed2
Programming languages (4)
Github contributions (5)
VivekBits2210/pykindler
Jul 2021 - Jul 2021
Automatic conversion of downloaded e-books to mobi format (and auto-upload to Kindle)
Contributions:3 releases, 95 commits, 16 PRs in 8 days
pythonmobiautomatic-conversionmobi-formatkindle
VivekBits2210/fine-tuning-large-language-models-on-limited-hardware
May 2023 - May 2024
NYU Tandon, ECE-GY 9143: High Performance Machine Learning, End Semester Project
Contributions:160 pushes, 1 branch in 1 year
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.