Apoorv Kulshreshtha

Lead ML Engineer

Mountain View, California, United States11 yrs 1 mo experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in Machine Learning and Deep Learning.
Led significant AI safety initiatives at Google.
Developed advanced conversational AI models.

Stackforce AI infers this person is a Machine Learning expert in AI-driven applications.

Contact

Skills

Core Skills

Machine LearningDeep LearningGenerative AiNatural Language Processing

Other Skills

AlgorithmsAmazon Web Services (AWS)Apache FlumeApache SparkCC++CUDAData StructuresEclipseElasticSearchGPGPUJavaMySQLParallel ProgrammingPython

About

My interests lie in the areas of Machine Learning, Deep Learning and NLP, as well as the Large Scale Distributed systems required for such data intensive applications.

Experience

Google

4 roles

Staff Software Engineer, Machine Learning

Promoted

Oct 2023 – Present · 2 yrs 5 mos

• Leading Gen AI Safety at YouTube

Machine LearningDeep Learning

Senior Software Engineer, Machine Learning

Apr 2021 – Oct 2023 · 2 yrs 6 mos

YouTube Trust and Safety
Started and leading Generative AI for Trust and Safety
Google Deepmind (previously Google Brain): LaMDA/Bard
TL for MultiLaMDA i.e. Multi-modal LaMDA
Previously TL for LaMDA Response Quality efforts. Led improvements in conversation Sensibleness, Specificity, Interestingness and multiple other characteristics
Sundar's Google I/O Keynote on LaMDA: https://youtu.be/Mlk888FiI8A?t=1075
LaMDA research paper: https://arxiv.org/abs/2201.08239
LaMDA Blog: https://www.blog.google/technology/ai/lamda

Machine LearningGenerative AI

Software Engineer III, Machine Learning

Promoted

Apr 2019 – Mar 2021 · 1 yr 11 mos

YouTube Trust and Safety
ML models for text and video understanding, deployed in production, at Google scale
Work spanning multiple verticals: Hate(ful) Speech, Threats, Minor Solicitation, Misinformation
Google Brain: Project Meena
Giant Language Modeling for Conversational AI
Led Meena Response Quality efforts, focussed on improving conversation Sensibleness, Specificity, Interestingness and multiple other characteristics
Meena Paper: https://arxiv.org/abs/2001.09977
Blog post: https://ai.googleblog.com/2020/01/towards-conversational-agent-that-can.html
Article: https://venturebeat.com/2020/01/28/meena-is-googles-attempt-at-making-true-conversational-ai/

Machine LearningDeep Learning

Software Engineer II, Machine Learning

Feb 2018 – Mar 2019 · 1 yr 1 mo

YouTube Trust and Safety
ML models for text and video understanding, deployed in production, at Google scale

Machine Learning

Columbia university in the city of new york

3 roles

Graduate Teaching Assistant

Sep 2017 – Dec 2017 · 3 mos · Greater New York City Area

Teaching Assistant for Fall 2017 Natural Language Processing course

Graduate Research Assistant

Sep 2017 – Dec 2017 · 3 mos · Greater New York City Area

Research Assistant at the Columbia NLP Lab

Graduate Student Researcher

Jan 2017 – May 2017 · 4 mos · Greater New York City Area

Worked with Prof. Kathleen McKeown on using Deep Learning and NLP for predicting scene changes in a novel
Created a neural machine translation model to classify discourse connectives as temporal/non-temporal based on whether the translation in the target language corresponded to temporal/non-temporal (The target language was chosen such that it had different translation of a particular discourse connective when used in temporal sense, as compared to when used in non-temporal sense)
Used distant supervision from French and Dutch to create weakly labeled data in English, for training, validation and testing neural machine translation model

Google

Software Engineering Intern

May 2017 – Aug 2017 · 3 mos · Mountain View, California

Indian institute of science (iisc)

Research Collaborator, Machine and Language Learning Lab

Sep 2014 – Sep 2016 · 2 yrs · Bengaluru Area, India

Worked on Deep Learning for temporal information extraction. The project dealt with predicting the publishing date of a document by using the temporal information of its content. The model developed, improved the accuracy of month prediction by 1.80 points, on the average month deviation scale, as compared to the existing models.

Deep LearningNatural Language Processing

Oracle india development center

Member of Technical Staff

Jul 2014 – Jul 2016 · 2 yrs · Bengaluru Area, India

Worked with Oracle Data Masking and Subsetting Team on product development and enhancement. Designed and implemented full database import/export functionality in integration with masking and sub-setting. Implemented improved memory management feature of choosing user defined tablespace during masking a database

Deep Learning

Flipkart.com

Intern

Jan 2014 – Jun 2014 · 5 mos · Bengaluru Area, India

Worked with Retail Analytics Team at Flipkart. Designed and implemented a system using Python, VBScript and QlikView to automate the creation of Category Management Dashboards for the following business verticals :
1. BGM (Books and General Merchandise),
2. CC-MT (Computers, Camera, Mobile and Tablets) and
3. Softline (Clothes and Fashion Accessories).