Kaushik B
Co-Founder & CTO at InXiteOut
Greater Kolkata Area India
Summary
🤩
Rockstar🎓
Top SchoolKaushik B is a hands-on Co-Founder & CTO with 20+ years of experience building AI-driven products and teams that turn data into measurable business value across EU, APAC, and the US. He leads InXiteOut’s AI practice, creating proprietary platforms like MEGHNAD for conversation intelligence and RIA, an agentic RAG tool, while architecting enterprise-grade GenAI and agentic AI solutions. His background spans Fortune 500s and startups, from leading the team that developed the world’s first 400GB high-speed microSD at Western Digital to winning an Excellence in Treasury award for an AI solution. A practitioner of both research and engineering, he has authored peer-reviewed papers, holds a US patent in storage, and contributes to notable open-source ML projects such as Ludwig and PyTorch Lightning improving hyperparameter tuning and training infrastructure. Known for scaling teams (100–200+ members) and operationalizing complex models, he blends strategy, product thinking, and deep technical expertise across ML, probabilistic models, and deep learning. Based in Greater Kolkata, he is available for speaking, research collaborations, and mentorship.
7 years of coding experience
12 years of employment as a software developer
University of TokyoB.E., Electronics & Telecommunication Engineering, B.E., Electronics & Telecommunication Engineering at Jadavpur UniversityMaster of Business Administration (MBA), Master of Business Administration (MBA) at Indian Institute of Management Bangalore
Github Skills (21)
data-visualizations10
webgl10
pytorch10
javascript10
python10
classification10
testing10
machine-learning10
data-visualisation10
pytorch-lightning10
hyperparameter-optimization10
deep-learning10
object-detection10
computer-vision10
data-visualization10
Programming languages (7)
Github contributions (5)
Lightning-AI/pytorch-lightning
Feb 2021 - Nov 2022
Pretrain, finetune ANY AI model of ANY size on multiple GPUs, TPUs with zero code changes.
Role in this project:
ML Engineer
ML EngineerContributions:8 releases, 1906 reviews, 267 commits in 1 year 9 months
Contributions summary:Kaushik's commits indicate a focus on modifying and testing various aspects of PyTorch Lightning's training loop and associated hooks, including changes to the execution order of hooks like `on_train_batch_end` and `on_epoch_end`. The code changes predominantly involve testing the correct functioning of these hooks within the training process, suggesting a role centered on verifying the behavior of the library's core training mechanisms. Their contributions include adjustments to logging behavior within callbacks and modifications to the data loading process, indicative of work related to the core training loop and supporting infrastructure.
pythonheadachespytorch-modelsdata-sciencehandling
Lightning-Universe/lightning-flash
Feb 2021 - Apr 2022
Your PyTorch AI Factory - Flash enables you to easily configure and run complex AI recipes for over 15 tasks across 7 data domains
Role in this project:ML Engineer
ML EngineerContributions:156 reviews, 24 commits, 42 PRs in 1 year 3 months
Contributions summary:Kaushik primarily contributed to the development of the `lightning-flash` repository, which focuses on building AI recipes using PyTorch. Their work included adding support for various backbones (MobilNet, VGG, DenseNet, ResNext) to the image classification and embedding models, enhancing the model's flexibility. They also refactored existing code, fixed import issues, and addressed bugs. Additionally, the user added detection task, incorporating a custom COCO dataset and fine-tuning capabilities.
torch-geometricicevisionconfigurepytorch-lightningdomains
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.