Vadim Smolyakov

Data Scientist

Microsoft

I am a Data Scientist at Microsoft. I was a PhD student in the Sensing, Learning, and Inference (SLI) lab at MIT CSAIL.

My research interests are in Bayesian inference, probabilistic modeling and deep learning in application to computer vision and natural language processing.

I completed my MASc degree specializing in wireless communications and BASc degree in Engineering Science at the University of Toronto

In the past, I interned as a Data Scientist at Microsoft, Rue Gilt Groupe and Shopify. I was also at Qualcomm and Max Linear as a systems research intern and a digital baseband designer.

News

12/2022: My book "Machine Learning Algorithms in Depth" is in Manning Early Access Preview
12/2021: We filed a patent on "Automatically forming and using a local network of smart edge devices"
07/2020: Our Responsible AI project was selected as a nominee for the best user value award
05/2020: I presented my work on hyper-personalization at MLADS (Microsoft internal AI conference)
11/2019: I joined Enterprise & Security DI R&D team at Microsoft in Redmond, WA
08/2019: I completed my data science internship at Microsoft working on intelligent edge

Projects

Research and Development

Responsible AI

Users in charge of information

V. Smolyakov, S. Ranganathan, S. Narayanan, J. Shotts, M. Todorova, E. Nouri, et al. "Responsible AI: Users in Charge of Information", 2020 [video]

Hyper-Personalization

Intent classification model for intelligent edge

V. Smolyakov, H. Liu, L. Shan, J. Liang, B. Nitta, "Intent Classification for Intelligent Edge", 2019 [slides]

Bayesian Nonparametric Models

Modeling driver behavior using HDP Split-Merge sampling algorithm

V. Smolyakov, J. Straub, S. Zheng, and J. W. Fisher III, "Bayesian Nonparametric Modeling of Driver Behavior using HDP Split-Merge Sampling Algorithm", 2018 [arXiv]

Minibatch Gibbs Sampling

Optimizing mini-batch size for Gibbs sampling algorithms on big data

V. Smolyakov, Q. Liu, J. W. Fisher III, "Adaptive Scan Gibbs Sampler for Large Scale Inference Problems", 2018 [arXiv] [slides]

Visual Search

Suggesting similar e-commerce products via object detection and locality-sensitive hashing

V. Smolyakov, R. Matthew, L. Zhang, "RDS Deep Learning Series: Visual Search and More", 2018 [slides]

Information Planning

Entropy and mutual information planning enables faster learning with fewer training examples.

V. Smolyakov, "Information Planning for Text Data", 2018 [arXiv]

Beamforming

mm-wave beam-search for MIMO wireless communication

V. Smolyakov, J. Ryu, S. Subramanian, "mm-wave beam search", 2013 [slides]

K-Best MIMO Detector

High-Throughput Digital Architecture for K-Best Multi-Input Multi-Output Detector

D. Patel, V. Smolyakov, M. Shabany, and P. G. Gulak, "VLSI Implementation of a WiMAX/LTE Compliant Low-Complexity High-Throughput Soft-Output K-Best MIMO Detector", 2010 [PDF]