Kartik Chaudhary

I am currently working as a Lead AI/ML Engineer at Google, where I design AI/ML based solutions/products leveraging the advancements in the field of Computer Vision, NLP, GenAI and Machine Learning.

I am passionate about learning and inventing new ways of improving ML and DL algorithms to solve real world problems more efficiently. Do checkout my blog on Artificial Intelligence - Drops of AI

Get in Touch

LinkedIn  /  Google Scholar  /  Twitter  / 
Github  /  Medium  /  Personal Blog

Quick links

News  |  Books  |  Publications  |  Patents  |  Articles

News

[2025]: Published a Google Blog on Optimizing image generation pipelines on Google Cloud!
[2024]: Published book on Generative Adversarial Networks titled The GAN Book!
[2024]: Published book on The Definitive Guide to Google Vertex AI with Packt!
[2023]: Got Promoted to Lead AI/ML Engineer at Google, May-2023!
[2022]: Filed a patent on Improving Wind Power production forecast using Image Processing techniques!
[2022]: Published Easter2.0 paper preprint on Arxiv and code on Github!
[2021]: Joined Google as a Senior AI/ML Engineer, Dec-2021!
[2021]: Awarded honorary title of Senior Inventor for several contributions to patents in 2020, Optum!

[2021]: Runner’s Up, Global Data Science Hackathon, Optum!
[2020]: Runner’s Up, Global Data Science Hackathon, Optum!

Books

If you are a ML practitioner who wants to learn about generative learning approaches and get expertise in Generative Adversarial Networks for generating high-quality and realistic content, this book is for you.

If you are a machine learning practitioner who wants to learn end-to-end ML solution development on Google Cloud Platform using MLOps best practices and tools offered by Google Vertex AI, this is the book for you.

Publications

Most of my research projects are related to computer vision, optical character recognition, image processing and video processing.

Optimizing image generation pipelines through hardware, code, and pipeline improvements boosts performance, cuts costs, and enhances user experience without sacrificing image quality.

Clinical Decision Support System (CDSS) is an important technology for the healthcare industry that analyzes data to help healthcare professionals make decisions related to patient care.

Easter2.0 is a small/fast convolutional model for the task of OCR/HTR that works well even when labelled data is limited.

Easter is a small/fast fully convolutional model for the task of OCR/HTR.

Patents

Dynamic triggering of augmented reality assistance mode functionalities

Automated systems and methods for identifying fields and regions of interest within a document image.

Dynamic detection of cross-document associations.

Dynamic detection of cross-document associations-II

Machine learning models for multi-risk-level disease spread forecasting.

Predictive document conversion.

Wind power production prediction using machine learning based image processing.

Training classification machine learning models with imbalanced training sets.

Articles

Understanding Audio data, Fourier Transform, FFT and Spectrogram features for a Speech Recognition System.

Explaining Reinforcement Learning to your next-door-neighbor.

Optimizers explained for training Neural Networks.

1D-CNN based Fully Convolutional Model for Handwriting Recognition.

Boosting your Sequence Generation Performance with ‘Beam Search + Language model’ decoding.

Deep Learning with PyTorch: Introduction.

Deep Learning with PyTorch: First Neural Network.

OpenCV: Introduction and Simple Tricks in Python.

Convolutional Denoising Autoencoders for image noise reduction.

Sound Wave Basics — Every Data Scientist must know before starting analysis on Audio Data.

Variational AutoEncoders and Image Generation with Keras.

this guy's webpage is awesome