Bhargav Chippada

AI Researcher

Mountain View, California, United States8 yrs 9 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in machine learning and video understanding.
Contributed to AV1 codec for predictive video compression.
Outstanding performance rating at Ola.

Stackforce AI infers this person is a Software Engineer specializing in Machine Learning and Geospatial Technologies.

Contact

Skills

Core Skills

Machine LearningDeep LearningVideo CompressionCloud ComputingData EngineeringGeospatial AnalysisSoftware DevelopmentMarketing AnalyticsMobile Development

Other Skills

Multimodal Deep LearningTransfer LearningDistillationC++Predictive Video CompressionCNNInter-frame CodingNeural NetworksSQLPythonSparkAzureData ProcessingGeospatial TechnologiesHidden Markov Model

About

I am a software engineer ML at Google, where I work on video understanding using LLMs and other ML techniques for Google Video Search - Shopping. I have a master's degree in computer science with a concentration in data science from the University of Massachusetts - Amherst, and a bachelor's degree in computer science from the Indian Institute of Technology, Bombay. With over six years of experience in software development, machine learning, geospatial technologies, and large-scale distributed systems, I have contributed to various research and development projects in the tech industry. Some of my notable achievements include working on the new version of AV1 codec, an open-sourced project by the Alliance for Open Media that uses predictive video compression to achieve higher compression rates and performance, and presenting a paper on generating humor and quotes using a controlled LSTM architecture at the Humanizing AI workshop in IJCAI. My goal is to leverage my skills and knowledge to create innovative solutions that enhance the user experience and benefit the society.

Experience

8 yrs 9 mos

Total Experience

2 yrs 2 mos

Average Tenure

5 yrs 5 mos

Current Experience

Google

2 roles

Machine Learning Engineer

Jan 2021 – Present · 5 yrs 5 mos · Mountain View, California, United States

Google Lens - Multimodal search
Improve google lens web results for multimodal queries (image + text)
Multimodal deep learning models, transfer learning, and distillation
Google Search - Shopping, Videos, and Human Voices.
Video and Web content understanding using LLMs and other ML techniques.

Machine LearningDeep LearningMultimodal Deep LearningTransfer LearningDistillation

Software Engineer Intern

May 2020 – Aug 2020 · 3 mos · Mountain View, California, United States

Worked on the new version of AV1 codec, an open-sourced project by the Alliance for Open Media (AOMedia) lead by Google which uses predictive video compression to achieve higher compression rates and performance. It's a perfect blend of my interest in CS, AI, and open-source.
Used CNN to restore the quality of degraded frames after its encoded using inter-frame coding technique
Used spatial separable convolutions to reduce the number of neural network parameters by 85%
Achieved -1.2% reduction in BD-rate for the same PSNR compared to the state of the art with just 10k parameters
Trained inter-frame CNN models on HD resolution video data for different quantization value buckets
Implemented a new method called Guided CNN Restoration and showed it's effectiveness for further gains

C++Predictive Video CompressionCNNInter-frame CodingNeural NetworksVideo Compression+1

University of massachusetts amherst

Graduate Grader

Aug 2020 – Dec 2020 · 4 mos · Amherst, Massachusetts, United States

Grader for the course COMPSCI 590Q: Quantum Information Systems

Microsoft

Software Engineer

Mar 2018 – Jul 2019 · 1 yr 4 mos · Greater Bengaluru Area

Bing - Local Search Team
Improve the relevance and ranking of Bing local search results
Improved the business listings ranking using an effective popularity measure for the businesses
Used signals like search counts, impressions, click counts, ratings, and reviews to predict a target
Azure Partner Experience Team
Migrated the Xbox360 purchases data processing pipeline to Azure cloud services without any downtime and issues
Integrated real-time purchases data into reports thereby cutting down the delay from 12 hours to 15 mins
Worked with the following technologies: Spark, Scala, SQL, Python, Azure, ADW

SQLPythonSparkAzureData ProcessingCloud Computing+1

Ola (ani technologies pvt. ltd)

Software Development Engineer

Jul 2016 – Mar 2018 · 1 yr 8 mos · Bengaluru Area, India

Location Intelligence Team
Problem Statement: Accurately track the position and direction of cab movement
on road network
Implementation:
Implemented Hidden Markov-Model based map matching algorithm for snapping
the noisy GPS updates of a cab to the right road and calculating its bearing
Built the algorithm on top of Open Street Routing Machine (OSRM) in C++ to
ensure speed while handling millions of requests per minute
Developed a routing server to route the requests to the appropriate distributed
city-wise cluster depending on the geo-location of a request
Improved the efficiency of Ola’s tracking system by 20%
Other Projects:
Developed a program to process the GPS traces of cabs, stored as polylines, to find missing road segments in map network data
Worked extensively with open street map data, time-dependent routing, route optimization, and mining geospatial data
Developed OlaMaps, an internal custom map tile server, and website for geo-visualizations
Developed a new technique for clustering and visualizing high-dimensional data by embracing an aspect of the dimensionality curse referred to as hub co-occurrence
Implemented a scalable community detection algorithm for weighted graphs in Spark
Achievement: Rated Outstanding (5/5) in the yearly appraisal

C++Geospatial TechnologiesHidden Markov ModelOpen Street Routing MachineGeospatial AnalysisSoftware Development

Amazon

Software Development Engineer Intern

May 2015 – Jul 2015 · 2 mos · Bengaluru, Karnataka, India

Problem Statement: Attribute the purchases of customers to marketing campaigns
Implementation:
Credited purchases of customers to different marketing channels using various
attribution models to identify effective marketing campaigns
Developed an easily extensible attribution model framework and implemented the
standard models like the last interaction, time decay, linear and position based

Attribution ModelsData AnalysisMarketing Analytics

Knit messaging

Android Developer

Mar 2014 – Jul 2014 · 4 mos · Mumbai Metropolitan Region

Joined this startup during its formative stage and co-developed the Android App for one-way messaging between a teacher and parents of their students for quick information dissemination.

Android DevelopmentMobile Development

Infosys

Software Engineering Intern

May 2008 – May 2008 · 0 mo · Greater Hyderabad Area

In 8th grade, I was one of the top 2 students in my city to receive an opportunity for an internship at Infosys as part of their "Catch Them Young" program. I learned the basics of object-oriented programming and developed a simple application for managing student registration.