Nayef Ahmed
Senior Software Engineer - IGML at Instagram
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolNayef Ahmed is a senior software engineer with nine years of experience building scalable ML and AI infrastructure, currently improving GPU efficiency and model training libraries at Instagram's IGML team. He previously contributed to PyTorch at Meta, helping build platforms for training LLMs, content understanding, and generative models, and has a track record of designing high-performance NLP and CV building blocks. A Mechatronics engineer with an AI option from the University of Waterloo, Nayef blends robotics intuition with backend systems expertise—evident from his RE2-based C++ regex tokenizer added to torchtext for production NLP workflows. He’s worked across research and product teams at top labs (Facebook AI, Lyft, Microsoft) and brings practical experience shipping systems that run at scale. Colocated in New York, he favors projects at the intersection of robotics, AI, and systems engineering.
9 years of coding experience
4 years of employment as a software developer
Charles III University of Madrid (Universidad Carlos III de Madrid)Bachelor of Applied Science - BASc, Honors Mechatronics Engineering, Option in Artificial Intelligence, Bachelor of Applied Science - BASc, Honors Mechatronics Engineering, Option in Artificial Intelligence at University of Waterloo
Github Skills (5)
nlp10
c-language10
cprogramming-language10
nlpjs10
natural-language-processing10
Programming languages (4)
Github contributions (5)
pytorch/text
May 2020 - Jan 2023
Models, data loaders and abstractions for language processing, powered by PyTorch
Role in this project:
Backend Developer
Backend DeveloperContributions:1 release, 443 reviews, 220 commits in 2 years 8 months
Contributions summary:Nayef primarily contributed to the `torchtext/text` repository by integrating regular expression functionalities. Their work included implementing a C++ class for regex tokenization using RE2 and adding the class to torchtext. The user implemented the `Sub` method which replaces the first match and added methods for applying regex operations and manipulating the stored tokens. This work involved building a custom class with the RE2 library for pattern matching, which was essential to the project and core to the natural language processing task.
nlppytorchloadersdeep-learningdataset
Nayef211/torcharrow
Mar 2022 - Jul 2022
A torch.Tensor-like DataFrame library supporting multiple execution runtimes and Arrow as a common memory format
Contributions:2 PRs, 73 pushes, 7 branches in 4 months
memorydataframeruntimestorcharrow
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.