Pankaj Rajan

ML Architect/Lead

Walmart Labs

January 2016 - November 2017

I was the Machine Learning Architect for Search at WalmartLabs and was responsible for architecting the next-generation search for Walmart. My team was behind one of the largest e-commerce search engines in the world that powers walmart.com/samsclub.com and enables millions of Walmart/Sams Customers to find what they desire. I also spent part of my time researching on deep learning applications for Information Retrieval, semantic understanding of products and queries and conversational shopping and technology evaulation. Previously, I incubated and lead a team for real-time query understanding which covers query classification, user intent understanding, brand identification, entity item relation, query disambiguation, query similarity, query augmentation, and expansion. My team explored problems at the intersection of statistical machine learning, probabilistic graphical models, deep learning.Our goal is to identify the explicit as well as the implicit intent of the user.

Architect and Research Engineering Lead

Stealth Mode Start -up , Stanford

June 2015 - December 2015

The architect of the IOT distributed system that can capture streams of sensor data from various sensors and enable the system to predict user position, motion, and presence. - Contributed to the development of Bayesian updated belief system for presence. - Contributed to the development of vision based gesture/action recognition system. - Developed data aggregation and analysis pipeline, which streams data from several properties for Large Scale Machine Learning. - Managed engineering roadmap and deliverables. - Worked with CEO for future roadmap, priorities and technical roadmap. - Streamlined Development using Agile methodologies, TDD and metric driven outcomes.

Sr. Software Engineer Architect and Lead Supervised ML Platform

Ayasdi Inc.

August 2014 - May 2015

Architected and Developed easy to use and highly scalable Supervised Machine Learning Platform on top of Unsupervised state of art Ayasdi TDA algorithm. -Designed and developed scalable statistical machine learning framework with (Random Forests /Decision Trees/Support Vector Machines /Feed Forward -Neural Network /AutoEncoders/Logistic Regression /Linear Regression /Localized Linear Regression (Variant) /Quadratic Discriminant Analysis classifiers) on top Topological Data Analysis. TDA is an unsupervised technique to identify pattern/ clusters in data. - Designed and Incubated Deep Neural Network Architecture (DNN, Deep Auto Encoders) for Ayasdi. -Developed POC NLP application on top of topological data analysis for insurance fraud detection. -Developed backend for distributed execution of ML processes, model storage, and model integrity. -Pioneered NLP and Text Analysis on Ayasdi Platform.

Sr. Applied Researcher, Machine Translation (Cognitive Computing Group)

eBay Inc

May 2013 - August 2014

Worked as a research software engineer for Machine Translation Team at eBay and contributed to design and development of highly scalable and high traffic Machine Translation Service.Developed Machine Translation Models For Real-Time Query Translation and Title Translation. Focus areas included: Machine Translation, Natural Language Processing (NLP), Statistical Machine Learning, Statistical Inference, Machine Learning. Key Contributions: Lead efforts for design and architecture of search back-end and MT integration to enable multi-lingual search queries on eBay.com as a part of GEO expansion initiative of eBay. Contributed significantly to Design and Development of Service-oriented Architecture for Machine Translation Service with lower tens of millisecond SLA. Contributed to design and development of in-house realtime machine translation system Contributed to prototyping, implementing and deploying various NLP components as a service. Primarily worked with: Java/ Spring/C++/Python/ MongoDB

Software Engineer II

Microsoft Corporation

January 2011 - May 2013

Worked as part of Dynamics Ax and Office for Mac: Designed and developed libraries for integration of Microsoft Excel to work with Dynamic Ax and Ax on The Web. As a Software Engineer for Microsoft's Apple Productivity Experience division (APEX) worked on email storage and backup migration code from custom DB to SQL lite.

Software Engineer II

AMD

June 2008 - January 2011

I was the engineering lead for developing production software for next generation (and only CPU+GPU) architecture at AMD. My team was responsible for the rewrite of the entire code base to move from STAR architecture (AMD Athlon) to more dynamic and easily scalable code base that can scale to multiple CPU architectures. Highlights of my tenure at AMD: * Developed framework that combined various signals and data to flag anomalies and reduced operator errors by up to 30%, resulting in significant saving for AMD manufacturing. * Lead of a global team responsible for delivering software for AMD manufacturing. • Designed, Architected and developed highly scalable and efficient C++ libraries that enable final fusing of AMD 32 and AMD 40 nm (GPU+CPU) chips. Millions of AMD CPU's were processed through the software developed by my team. I was also involved as a developer for AMD 45 nm and AMD 65 nm products. I have contributed to AMD Phenom II, AMD Athlon II, AMD Turion II, AMD Neo, AMD 6 and 12 core Opteron code.

Texas A&M University, College Station

Master of Science, Computer Science

2007 - 2009

National Institute of Technology, Rourkela

B.Tech, Mechanical Engineering

2002 - 2006