Neeraj Baji

AI Researcher

Seattle, Washington, United States16 yrs 2 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Led a team of 10 scientists at Amazon.
Developed scalable ad systems with significant budget management.
Expert in machine learning and deep learning applications.

Stackforce AI infers this person is a highly skilled professional in AdTech and Mobile Technology with expertise in machine learning.

Contact

Skills

Core Skills

Machine LearningAdtechDeep LearningMarketing PersonalizationRecommender SystemsMobile DevelopmentAcoustic Processing

Other Skills

AWSAcousticAcoustic AlgorithmsAcoustic Echo CancellationAirflowAlgorithmsAndroidAndroid DevelopmentApache CassandraApache KafkaApache SparkAudio ProcessingCDSPDigital Signal Processing

About

Science leader. Adept at building state of the art IP that works at scale across speech processing, recommender systems and AdTech.

Experience

Amazon

4 roles

Applied Science Manager

Sep 2022 – Present · 3 yrs 6 mos

Developed an extensible system that could scale from 5 campaigns across 2 advertisers, supporting $300k of Ad Spend in 2021 to 60 advertisers, 1000 campaigns YTD with $95MM in Ad Spend in 2023.
Grew the team from a single scientist (myself!) to 10 scientists; working with recruiters to get the best scientists to the team and actively mentoring to help them develop.
Leveraging best in class Open Source LLMs to power automated audience and campaign insights for streaming media advertisers

Machine LearningAdTechScalability

Senior Applied Scientist

Promoted

May 2020 – Present · 5 yrs 10 mos

Building and scaling Amazon Advertising across media verticals
Building an End to end machine learning based product spanning audience targeting, bidding optimization and creative optimization; targeted towards games and entertainment advertisers.
Set up foundational tech for the product spanning data onboarding, featurization, ML models and ML OPS. Implemented using Pyspark, XgBoost, PyTorch , HuggingFace and Airflow.

Machine LearningPysparkXgBoostPyTorchHuggingFaceAirflow+1

Senior Applied Scientist

Jun 2019 – May 2020 · 11 mos

Outbound marketing personalization at Amazon scale

Applied Scientist

Dec 2016 – Jun 2019 · 2 yrs 6 mos

Cross-channel marketing personalization.
Deep learning to personalize marketing emails, push notifications and sms at Amazon scale. Models developed range from gbm's, feed forward networks, neural collaborative filtering models, sequence models, etc.
Explore exploit models using Thompson sampling.
End to end model training and deployment workflow using AWS data pipelines, spark-emr clusters, gpu instances and S3. Handling billions of messages and millions of customers world wide.
Tools : Keras, tensorflow, scikit, Apache spark, EMR, AWS data pipelines.

Deep LearningAWSApache SparkMarketing Personalization

Ittiam systems pvt ltd

4 roles

Lead Engineer, Corporate R&D

Sep 2015 – Nov 2016 · 1 yr 2 mos

Applied research:
Deep learning for vision. Convnets, Tensorflow,Keras and more.
Lambda architecture based recommender using Apache Kafka, Apache Cassandra and Apache Spark ensuring fault tolerance and scalability.
Build a state of the art, scalable recommender system from scratch for the online video space.Tools being used include NLTK, Scikitlearn, Apache Spark, Kafka, HDFS and Cassandra.
Topic modeling, online learning for product differentiation. Solve cold start for new users and new items using clustering and an explore-exploit paradigm.
Handle deployment on the Google Cloud Platform.

Deep LearningApache KafkaApache CassandraApache SparkRecommender Systems

Senior Engineer, Corporate R&D

Jun 2015 – Sep 2015 · 3 mos

Build a state of the art, scalable recommender system from scratch for the online video space.Tools being used include NLTK, Scikitlearn, Apache Spark, Kafka, HDFS and Cassandra.

Senior Engineer, Mobile Acoustics

Promoted

Oct 2012 – May 2015 · 2 yrs 7 mos

Android app development. Making products out of state of the art IP.
AEC, NS, AGC, etc tuning for various mobile devices and tablets.
Development of android based mobile apps for Acoustic IP demonstrations.
Performance improvement of specific IPs like Linear array beamforming & DMNC.
Acoustic IP performance validation.
IP demo bring up on various platforms.
Integral part of the team that demoed Ittiam's "Linear Microphone Array Beamforming" at CES 2013.

Engineer

Feb 2012 – Oct 2012 · 8 mos

Development of Acoustic algorithms.
Integration of acoustic IP in the Android audio framework.
Android app development.

Android DevelopmentAcoustic AlgorithmsMobile DevelopmentAcoustic Processing

Hackathons

Weekend Hacker

Feb 2015 – Mar 2016 · 1 yr 1 mo

Ola Hackathon - 4
th prize:
Developed an Android application that allows users to reserve parking spaces and
apartments, buildings with extra parking space to monetize it. Used the Hungarian algorithm at the backend
for optimal demand matching and minimize conflicts.
Sequoia Hack 2015:
Among the few teams that participated in the Sequoia Hack under the AI theme.
Developed a chatbot that used medical data from WebMD and allowed users to get a first level diagnosis by
describing their symptoms in natural language.
AppLift Hack 2015 - 2nd Prize:
Given anonymized ad exchange bid data, developed a model and a
corresponding R shiny app that allowed AppLift to generate the optimal ad placement strategy to maximize
click through rate given a specific ad budget and given customer’s inclination towards diversified ad
allocation.

Acoustic AlgorithmsAndroid DevelopmentAcoustic ProcessingMobile Development

Robert bosch engineering and business solutions ltd.

Engineer

Jul 2011 – Jan 2012 · 6 mos · Bengaluru Area, India

Tcs

Assistant Systems Engineer

Dec 2009 – Aug 2011 · 1 yr 8 mos · Bangalore

DSP engineer for TCS-Aeroflex Test Solutions UK
Implementation of TD-SCDMA physical layer for Aeroflex 7100 Base Station Emulator on TI C64X dsp. Implemented PHY layer algorithms like Rate Matching,Reed Muller codec, subframe segmentation,physical channel mapping & midamble code generation (25.221 - 25.224 3GPP specs).
Android developer for TCS-Firethorn(Qualcomm).
Developed a POC for push notifications for Android applications based on Android 1.6 to 2.1 using IBM's MQTT protocol.
J2EE developer for TCS-Qualcomm.
Developed an engineering web app allowing Toyota engineers to analyze logged parameters on Portable Hybrid Electric Vehicles depending on specific date ranges. Technologies used included XStream, Spring 3.0, JSF 1.1