Mountain View, California, United States
Contact Info
11K followers
500+ connections
About
Articles by Deepak
-
2017 Grace Hopper Celebration of Women in Computing
2017 Grace Hopper Celebration of Women in Computing
By Deepak Agarwal
Activity
-
"Makeathon is my favorite thing about Pinterest. This is a place where everyone is encouraged to build and contribute to the product." 📌 - A'Dream…
"Makeathon is my favorite thing about Pinterest. This is a place where everyone is encouraged to build and contribute to the product." 📌 - A'Dream…
Liked by Deepak Agarwal
-
In just one year, Pinterest went from no generative AI capabilities to a fully built out platform that has transformed productivity and product…
In just one year, Pinterest went from no generative AI capabilities to a fully built out platform that has transformed productivity and product…
Liked by Deepak Agarwal
-
If you are looking for something to read this weekend, I am happy to share that Chapter 7 on instruction finetuning LLMs is now finally live on the…
If you are looking for something to read this weekend, I am happy to share that Chapter 7 on instruction finetuning LLMs is now finally live on the…
Liked by Deepak Agarwal
Experience & Education
Licenses & Certifications
Publications
-
Scout: A Point of Presence Recommendation System Using Real User Monitoring Data
Passive and Active Measurements Conference
This paper describes, Scout, a statistical modeling driven approach to automatically recommend new Point of Presence (PoP) centers for web sites. PoPs help reduce a website’s page download time dramatically. However, where to build the new PoP centers given the current assets of existing ones is a problem that has rarely been studied in a quantitative and principled way before; it was mainly done through empirical studies or through applying industry experience and intuitions. In this paper, we…
This paper describes, Scout, a statistical modeling driven approach to automatically recommend new Point of Presence (PoP) centers for web sites. PoPs help reduce a website’s page download time dramatically. However, where to build the new PoP centers given the current assets of existing ones is a problem that has rarely been studied in a quantitative and principled way before; it was mainly done through empirical studies or through applying industry experience and intuitions. In this paper, we propose a novel approach that estimates the impact of the PoP centers by building a statistical model using the real user monitoring data collected by the web sites and recommend the next PoPs to build. We also consider the problem of recommending PoPs using other metrics such as user’s number of page views. We show empirically that our approach works well, by experiments that use real data collected from millions of user visits in a major social network site.
-
Statistical Methods for Recommender Systems
Cambridge University Press
This book summarizes research done for a period of 5 years when I was at Yahoo! Research. Much of the work was motivated by our project of personalizing the Yahoo! front page. Many of the methods discussed are deployed both at yahoo! and linkedin.
Other authorsSee publication -
Automatic Ad Format Selection via Contextual Bandits
Conference on Information and Knowledge Management (CIKM)
Visual design plays an important role in online display advertising: changing the layout of an online ad can increase or decrease its effectiveness, measured in terms of click-through rate (CTR) or total revenue. The decision of which layout to use for an ad involves a trade-off: using a layout provides feedback about its effectiveness (exploration), but collecting that feedback requires sacrificing the immediate reward of using a layout we already know is effective (exploitation). To balance…
Visual design plays an important role in online display advertising: changing the layout of an online ad can increase or decrease its effectiveness, measured in terms of click-through rate (CTR) or total revenue. The decision of which layout to use for an ad involves a trade-off: using a layout provides feedback about its effectiveness (exploration), but collecting that feedback requires sacrificing the immediate reward of using a layout we already know is effective (exploitation). To balance exploration with exploitation, we pose automatic layout selection as a contextual bandit problem.
There are many bandit algorithms, each generating a policy which must be evaluated. It is impractical to test each policy on live traffic. However, we have found that offline replay (a.k.a. exploration scavenging) can be adapted to provide an accurate estimator for the performance of ad layout policies at Linkedin, using only historical data about the effectiveness of layouts. We describe the development of our offline replayer, and benchmark a number of common bandit algorithms.Other authorsSee publication
Patents
-
Determining a Churn Probability for a Subscriber of a Social Network Service
Issued US
-
Determining user preference of items based on user ratings and user features
Issued US
A set of item-item affinities for a plurality of items is determined based on collaborative-filtering techniques. A set of an item's nearest neighbor items based on the set of item-item affinities is determined. A set of user feature-item affinities for the plurality of items and a set of user features is determined based on least squared regression. A set of a user feature's nearest neighbor items is determined based in part on the set of user feature-item affinities. Compatible affinity…
A set of item-item affinities for a plurality of items is determined based on collaborative-filtering techniques. A set of an item's nearest neighbor items based on the set of item-item affinities is determined. A set of user feature-item affinities for the plurality of items and a set of user features is determined based on least squared regression. A set of a user feature's nearest neighbor items is determined based in part on the set of user feature-item affinities. Compatible affinity weights for nearest neighbor items of each item and each user feature are determined and stored. Based on user features of a particular user and items a particular user has consumed, a set of nearest neighbor items comprising nearest neighbor items for user features of the user and items the user has consumed are identified as a set of candidate items, and affinity scores of candidate items are determined. Based at least in part on the affinity scores, a candidate item from the set of candidate items is recommended to the user.
-
Available from uspto as specified below
US Patents (filed and issued)
http://patft.uspto.gov/
Use advanced query: "Agarwal-Deepak$ and CA"
~27 patents issued, several pending. -
Determining a School Rank Utilizing Perturbed Data Sets
US
Honors & Awards
-
Fellow of the American Statistical Association
-
-
Member, Board of Directors, SIGKDD
-
Recommendations received
2 people have recommended Deepak
Join now to viewMore activity by Deepak
-
In a compelling session at VB Transform, Pinterest leaders shared transformative and practical strategies for embracing generative AI at scale within…
In a compelling session at VB Transform, Pinterest leaders shared transformative and practical strategies for embracing generative AI at scale within…
Liked by Deepak Agarwal
-
“It takes a village.” It was a wonderful day discussing how to build high-performing cross-functional product teams that design and build AI. Thanks…
“It takes a village.” It was a wonderful day discussing how to build high-performing cross-functional product teams that design and build AI. Thanks…
Liked by Deepak Agarwal
-
MobileLLM: nice paper from @AIatMeta about running sub-billion LLMs on smartphones and other edge devices. TL;DR: more depth, not width; shared…
MobileLLM: nice paper from @AIatMeta about running sub-billion LLMs on smartphones and other edge devices. TL;DR: more depth, not width; shared…
Liked by Deepak Agarwal
-
Nobody does "bring your kids to work day" quite like Pinterest! 📌 Yesterday, we were honored to welcome so many esteemed leaders, including…
Nobody does "bring your kids to work day" quite like Pinterest! 📌 Yesterday, we were honored to welcome so many esteemed leaders, including…
Liked by Deepak Agarwal
-
Autobound ranked in the top 5 for product of the day! Really appreciate everyone's support. Also about to close the books on our FIRST profitable…
Autobound ranked in the top 5 for product of the day! Really appreciate everyone's support. Also about to close the books on our FIRST profitable…
Liked by Deepak Agarwal
-
Interested in working on Trust & Safety at Pinterest?! My team is hiring four T&S PM roles covering a wide range of topics, including content…
Interested in working on Trust & Safety at Pinterest?! My team is hiring four T&S PM roles covering a wide range of topics, including content…
Liked by Deepak Agarwal
-
This week I spoke at the retail conference K5 - Future Retail in Berlin where I shared more about the new era of Performance at Pinterest and why…
This week I spoke at the retail conference K5 - Future Retail in Berlin where I shared more about the new era of Performance at Pinterest and why…
Liked by Deepak Agarwal
-
Thank you for the very warm welcome, Ashok Srivastava! I am excited to be part of your team at Intuit. I look forward to working closely with team…
Thank you for the very warm welcome, Ashok Srivastava! I am excited to be part of your team at Intuit. I look forward to working closely with team…
Liked by Deepak Agarwal
-
Join us for "Synergy in Synthesis: Forging the Future of AI with Cross-Functional Expertise" at VentureBeat Transform. Discover the power of…
Join us for "Synergy in Synthesis: Forging the Future of AI with Cross-Functional Expertise" at VentureBeat Transform. Discover the power of…
Liked by Deepak Agarwal
-
Excited to share that we've launched a new Board Sharing feature that allows people to share their Boards with the world in a video format!…
Excited to share that we've launched a new Board Sharing feature that allows people to share their Boards with the world in a video format!…
Liked by Deepak Agarwal
-
Applications are now open for our next cohort of Pinterest Products Apprentices! This is an incredible opportunity for candidates from…
Applications are now open for our next cohort of Pinterest Products Apprentices! This is an incredible opportunity for candidates from…
Liked by Deepak Agarwal
-
Part 2 is here! 🥳 Learn about ‘Ray Infrastructure at Pinterest’ written by Chia-Wei Chen, Jiun-Yu Lee, Alex Wang, Saurabh Vishwas Joshi, Karthik A…
Part 2 is here! 🥳 Learn about ‘Ray Infrastructure at Pinterest’ written by Chia-Wei Chen, Jiun-Yu Lee, Alex Wang, Saurabh Vishwas Joshi, Karthik A…
Liked by Deepak Agarwal
-
Pinterest will be at CVPR this week, stop by our booth to hear more about our latest work. Looking forward to seeing folks in person!
Pinterest will be at CVPR this week, stop by our booth to hear more about our latest work. Looking forward to seeing folks in person!
Liked by Deepak Agarwal
-
I'm looking for an amazing L16/L6 Technical Program Manager with prior Trust & Safety experience to join my team! If you know someone who would be a…
I'm looking for an amazing L16/L6 Technical Program Manager with prior Trust & Safety experience to join my team! If you know someone who would be a…
Liked by Deepak Agarwal
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore MoreOthers named Deepak Agarwal in United States
-
Deepak Agarwal
-
Deepak Agarwal
Head of Finance | C.A. in Pharmaceuticals
-
Deepak Agarwal
-
Deepak Agarwal
Software Development Engineer III @ Amazon Web Services | Software Development
65 others named Deepak Agarwal in United States are on LinkedIn
See others named Deepak Agarwal