Vijay Kalmath
Machine Learning Research Engineer at Scale AI
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolVijay Kalmath is a Machine Learning Research Engineer with nine years of experience merging deep learning research and production ML systems, currently working at Scale AI after completing an MS in Data Science at Columbia. He has hands-on expertise fine-tuning and optimizing LLMs (QLORA/LORA/PEFT), accelerating BERT inference with ONNX and quantization, and building reproducible model pipelines and determinism frameworks for both TensorFlow and PyTorch. At Lexalytics he boosted multilingual key-phrase extraction to 0.85 F1 and generated synthetic datasets for demos without using real customer data, while at Columbia he maintained the AIModelShare package on conda-forge with tens of thousands of downloads. His open-source contributions to the TextAttack framework show practical skills in adversarial NLP, debugging gradient issues and integrating tooling like Weights & Biases. Equally comfortable in DevOps and networking contexts from a prior Cisco career, he combines ML research rigor with production engineering and team leadership, having mentored and hired across functions.
9 years of coding experience
5 years of employment as a software developer
Bachelor of Engineering, Electronics and Communications Engineering, 9.37/10.0, Bachelor of Engineering, Electronics and Communications Engineering, 9.37/10.0 at B. M. S. College of EngineeringMaster of Science - MS, Data Science, 3.8/4.0, Master of Science - MS, Data Science, 3.8/4.0 at Columbia University in the City of New York
Stackoverflow
178reputation
7kreached
13answers
8questions
Github Skills (15)
pytorch10
machine-learning10
nlp10
adversarial-attacks10
python10
weight9
data-augmentation9
weighting9
deep-learning6
csv6
tensorflow6
tflearn6
append6
database6
macos-sierra6
Programming languages (3)
Github contributions (5)
QData/TextAttack
May 2022 - Jul 2022
TextAttack 🐙 is a Python framework for adversarial attacks, data augmentation, and model training in NLP https://textattack.readthedocs.io/en/master/
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 38 commits, 16 PRs in 1 month
Contributions summary:Vijay primarily contributed to the improvement and maintenance of the TextAttack framework, a Python library for adversarial attacks in NLP. Their work involved fixing bugs related to the integration of logging libraries like Weights & Biases, specifically addressing issues with data types and project initialization. They also focused on correcting gradient calculations within the PyTorch model wrapper for WordCNN, optimizing code related to dataset handling and attack recipe configurations. Furthermore, the user demonstrated expertise in handling errors and edge cases within the library's attack process.
pythonadversarial-machine-learningadversarialmodel-trainingtraining
VijayKalmath/vijaykalmath.github.io
Jul 2022 - Dec 2023
Personal Portfolio
Contributions:2 PRs, 113 pushes, 2 branches in 1 year 4 months
cssjavascriptreactreactjs
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.