Soumik Rakshit
Machine Learning Engineer at Google developers
Bengaluru, Karnataka, India
Summary
🤩
Rockstar🎓
Top SchoolSoumik Rakshit is a Machine Learning Engineer with nine years of experience building and deploying computer vision and AI tooling, currently working at Zomato and recognized as a Google Developers Expert. He has strong research-to-production chops—contributing to high-profile open-source projects like wandb (enhancing experiment tracking for Keras/Ultralytics) and Keras/Keras-io (implementing DeepLabV3+ and Keras Core compatibility). His work spans HDR imaging, 3D point cloud segmentation, real-time rendering and graphics engines, and he’s actively exploring low-level acceleration (CUDA/OpenCL, Vulkan, DirectX) and JAX. Soumik blends research rigor with practical engineering: examples include StyleGAN-NADA notebooks, post-training optimizations for foundation models, and migration of TensorFlow logic to Keras Core. Outside of code he writes horror stories and streams creative content, reflecting a curiosity that fuels both deep technical work and accessible educational material.
9 years of coding experience
9 years of employment as a software developer
Bachelor of Technology (B.Tech.), Computer Science And Engineering, Bachelor of Technology (B.Tech.), Computer Science And Engineering at Kalinga Institute of Industrial TechnologyCBSE, 12th Standard, Computer Science, Physics, Chemistry, Mathematics, CBSE, 12th Standard, Computer Science, Physics, Chemistry, Mathematics at Bholananda National VidyalayaICSE, 10th Standard, General Science, Computer Applications, ICSE, 10th Standard, General Science, Computer Applications at Modern English Academy
English, Hindi, Bengali, German
Stackoverflow
879reputation
370kreached
9answers
10questions
Github Skills (32)
pytorch10
python10
semantic-segmentation10
machine-learning10
stylegan10
tracks10
keras10
google-colab10
deeplearning-ai10
deep-learning10
tensorflow10
resnet10
ai10
wandb10
google-colaboratory10
Programming languages (10)
Github contributions (5)
wandb/examples
Jun 2022 - Jan 2023
Example deep learning projects that use wandb's features.
Role in this project:
ML Engineer
ML EngineerContributions:9 reviews, 37 commits, 34 PRs in 6 months
Contributions summary:Soumik contributed a comprehensive notebook detailing the implementation and training of a StyleGAN-NADA model. This involved setting up the necessary libraries, configuring the model type, fetching pretrained models from WandB artifacts, and defining the training process. The notebook covers model training with the ability to specify source and target classes.
pytorchdeep-learningdeep-learning-examplemachine-learningwandb
keras-team/keras-io
Sep 2021 - Feb 2022
Keras documentation, hosted live at keras.io
Role in this project:ML Engineer
ML EngineerContributions:41 reviews, 6 commits, 11 PRs in 5 months
Contributions summary:Soumik primarily contributed to the implementation and refinement of a DeepLabV3+ model for multiclass semantic segmentation within the Keras ecosystem. Their work involved adding, updating, and modifying the example code, including incorporating the ResNet101 backbone and generating required files. The commits suggest an active role in developing and testing the model, along with making iterative improvements. The user also added validation, indicating a focus on model performance.
deep-learninghostedmachine-learningtensorflowkeras
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.