AI Scientist | Roman Kazinnik

Active Learning: End-to-End solution for Deduplication

Why we need TensorFlow Extended (TFX) - and how to get it in 3 steps

Distributed Hyperparameter Search in Kubeflow/Kubernetes: Keras Tuner vs. Katib

From Data Pipeline to Machine Learning Architecture in 3 simple pictures

An Ultimate Guide to Migrating Your Machine Learning to Kubeflow: Excelling in Production in 4 steps

Machine Learning as a Flow: Kubeflow vs. Metaflow

Machine Learning Distributed: Ring-Reduce vs. All-Reduce

Convolutional Neural Networks for NLP

Automate Machine Learning as a Flow: Kubeflow

Beyond Supervised ML: learning latent distributions

Convolutional Neural Networks: Recipe to Find Optimal Architecture and Avoid Redundancy

Beyond DAG: Introducing Cyclic Graphs in Neural Networks

Approximation theory, AI, and how they actually go together (talk at UNCC)

Roman Kazinnik -

Mathematics Ph.D. Machine Learning. Infrastructure.

For nearly 20 years, I have remained committed to the advancement of science, algorithms and modeling.

Currently, I lead a development of a broad range of forecasting and learning algorithms, with applications ranging from finance and marketing to machine learning infrastructure.

AI, Math and Data Science blog

Get in touch // Tel: 832-359-1379 // roman.kazinnik@gmail.com

ML, ML Infra, Math, and in Between.

sign up for my blog and newsletter