About
Activity
7K followers
-
Hitesh Kumar liked thisHitesh Kumar liked thisWe've given away over 6,000 books in the last 3 weeks. A couple weeks ago, I posted here on LinkedIn about a book we've been giving out at SF AI meetups. It's called Principles of Building AI Agents, and it's become popular among developers starting to build with AI, as well as their PMs and CEOs. One engineer said that he started reading after a person he didn't know gave him one on a bus. A current YC founder called it "the most popular book in SF right now." I expected to get a few dozen commenters. Instead, over 3,000 people have dropped us a line. My LinkedIn DMs went crazy, and our team's been sending out copy after copy. And if you want one, just comment "BOOK" and we'll send it over!
-
Hitesh Kumar liked thisHitesh Kumar liked thisTIME just named Partiful one of the 100 Most Influential Companies of 2025 🎉. It’s a pinch-me moment to see our small, scrappy team (under 30 people!) recognized alongside some of the biggest companies in the world. We’re honored to be included, and grateful to our global community of party people: this is a testament to your creativity and commitment to making the real world feel vibrant, joyful, and worth showing up for. Innovation in consumer apps is rare these days. Most apps we use were designed over 15 years ago(!)—and they haven’t changed much since then. Partiful is one of the only new consumer apps that has hit the scale of millions of users, with real staying power. Our phones, by and large, feel stuck in the past. And what about the real world? There are only so many minutes we can, or should, spend on a phone. New apps promise to solve loneliness, but they’re only offering us more screen time. With Partiful, we’ve certainly designed a beautiful app: a place for play and joy and delight, a place to save you time and stress (plan events in seconds—literally). But fundamentally, Partiful is designed to get you off of your screen and actually hanging out in the real world. Powerful things happen when we invest in our time together IRL. A single event planned on Partiful can lead to dozens of friendships, or even whole new friend groups. It can be the place for a meet-cute that turns into a relationship, an engagement, a marriage. It can be the start of a creative collaboration, a grassroots movement, or a community that changes someone’s life. Partiful breaks the doomscrolling loop and tethers you back to what makes you human: hugging your friends, high-fiving a stranger, locking eyes with the cutie across the room, laughing with your childhood BFF, dancing with your future soulmate, letting a DJ save your life. Partiful powers the moments that turn into lifelong memories. It feels cliché to say that we’re still just getting started, but there’s truly so much more we have left to build, to support the billions of people craving more authentic connection in their life. We’re hiring across almost all teams right now—if you or someone you know should be on this mission with us, please check out our open roles! https://lnkd.in/engj9XaR https://lnkd.in/eNJNTsSv
-
Hitesh Kumar liked thisHitesh Kumar liked thisX = X - 0.65 * XFidelity has marked down the value of Twitter/X by 65%Fidelity has marked down the value of Twitter/X by 65%
-
Hitesh Kumar liked thisHitesh Kumar liked thisLoved the chance to partner with Anthony Ayers and the city of Cottage City to deploy Obvio Inc's first product: the world's first stop sign detection trailer. Looking forward to partnering with more cities to keep our roads safe! Learn more at http://obvio.ai
-
Hitesh Kumar liked thisHitesh Kumar liked thisI feel delighted to inform that yesterday I completed an Online course by Stéphane Maarek on streaming Data using Kafka on Udemy . I did hands on creating Producers and Consumers in Java and my fundamentals became really strong. So what is kafka - Kafka is a distributed streaming platform that allows you to publish and subscribe to streams of records, similar to a message queue or enterprise messaging system. At its core, Kafka consists of three main components: Producers: Applications that write data to Kafka topics. Topics: A stream of records, which is a category or feed name to which records are published. Consumers: Applications that read data from Kafka topics. Kafka is designed to be highly scalable, fault-tolerant, and performant. It achieves this by using a distributed architecture with multiple brokers (servers) that can handle large amounts of data and provide high availability. Each Kafka topic is divided into multiple partitions, and each partition is replicated across multiple brokers to ensure fault tolerance. Producers can write data to any partition, and consumers can read data from any partition, which makes it easy to scale both producers and consumers horizontally. In addition, Kafka provides a publish-subscribe model that allows multiple consumers to read from the same topic. This makes it easy to build real-time data pipelines, stream processing applications, and event-driven architectures. Credit - Thanks Stéphane Maarek for creating such an insightful course. #data #streaming #architecture #kafkastreams #dataengineering
-
Hitesh Kumar liked thisHitesh Kumar liked thisAirbnb is looking for a Product Manager for their Trust&Safety org. As part of the team, you will have the opportunity to work on exciting set problems related to product & machine learning for keeping the platform safe. I recently joined the team and it's awesome. Apply below if you are interested https://lnkd.in/gdY3wuAT #machinelearning #hiring
-
Hitesh Kumar liked thisHitesh Kumar liked thisSuper excited to share that our work, led by Adam Leon, for community building of group chats on LinkedIn messaging is now published on the engineering blog! #linkedin #engineering #communityCommunity building recommender for group chats in LinkedIn MessagingCommunity building recommender for group chats in LinkedIn Messaging
-
Hitesh Kumar liked thisHitesh Kumar liked thisI recently joined Socure on the mission to eliminate identity frauds online. We have the best solutions in the market and have many high-profile customers in banking, credit card, FinTech and gaming industries. Our business is growing very fast and we are hiring aggressively across functions. Check out career opportunities at Socure https://lnkd.in/eYhQUYXT. I am #hiring Android, iOS and frontend engineers across levels. Come join me to eliminate 100% identity frauds!Career Opportunities & Current Job Openings | SocureCareer Opportunities & Current Job Openings | Socure
Experience & Education
Licenses & Certifications
Courses
-
Algorithms
-
-
Artificial Intelligence
-
-
Artificial Neural Networks
-
-
Cloud Computing
-
-
Compilers
-
-
Computer Networks
-
-
Information Retrieval and Extraction
-
-
Natural Language Processing
-
-
Pattern Recognition
-
-
Principles of Information Security
-
-
Software Engineering
-
Projects
-
Text Classification
- Present
In this I used a naive bayes based classifier for newsgroup text classification and tried out its various probabilistic variants. After that I used a large corpus of unlabelled documents with some labelled documents and combined EM algorithm with Naive Bayes to improve accuracy. Then used SVM and logistic regression in place of naive bayes. Used Java, Eclipse, R and Mahout,
-
Building a scalable search engine.
-
Build an index, retrieval and ranking module (using vector space model) on the wikipedia dataset. Used Java, Lucene and Eclipse.
-
Neon Code Integration
-
Optimized Hardware accelarated WMV codec and integrated neon optimized Software codec in Android. Also measured the resolution that can be supported. Used C++
-
Compiler for a subset of C Language
-
Build a compiler for a subset of C language as a part of compilers course. Used Flex,Bison and C++.
-
TreeBanking Tools Development
-
Developed tools like parser, POS tagger, stemmer(POS tagger) required for ongoing treebanking project of LTRC, IIIT-H. Used Python and Java Swing.
Test Scores
-
GRE
Score: 320/340
Languages
-
English
Full professional proficiency
-
Hindi
Full professional proficiency
-
Punjabi
Limited working proficiency
Recommendations received
1 person has recommended Hitesh
Join now to viewOther similar profiles
Explore more posts
-
Guy Lebanon
Cool recent paper from our team (Instagram Ads Ranking) on customizing the learning process of attention layers to preserve mutual information with minority cohorts while improving global performance. This could be quite relevant to other industry groups working on large scale recommendation systems. https://lnkd.in/gyfVWbS5
1 Comment -
Alison Smith
You might think model performance is the most important factor in moving a system to production. Waymo, however, judges how mature a project is by how mature its evaluations are. With over 220 million rider-only miles and 17x fewer serious injuries than human drivers, we probably can learn something from their approach. Manasi Joshi, Waymo's Director of Engineering, called it eval-forced development at VB Transform. She described testing as something that runs through the whole build, and not just a gate before launch. Unlike in traditional software where writing and verifying tests are relatively cheap and straightforward, Physical AI systems make judgment calls where answers aren't written down and codified. Building an eval set therefore becomes an exercise in extracting that expertise and making it machine-readable. If you think about it, the eval set actually becomes the asset that compounds, not the model. With EO14402 pushing DoW contracts over $100M toward firm-fixed-price, precise outcome definition becomes even more necessary for AI systems, which will sharpen what "good" means for evals too. https://lnkd.in/eV24Czsf #AI #EnterpriseAI #MachineLearning #AIAgents #DefenseTech
1 Comment -
Behind the CMO
Evan Spiegel is pulling resources out of Snap's engineering org and moving them to distribution. More than two-thirds of Snap's new code is written by AI now. A service that used to need a team takes "half a person's time." So when the CEO looks at where the advantage went, he doesn't point at the product. He points at distribution, and he's moving the budget to match. Read that as a CMO. A public-company CEO just made the case for marketing better than most marketing teams make it for themselves. The part that should keep you up: distribution is the new moat, and it's also getting eaten. Rented reach isn't a moat. It's a bill that climbs with demand. The defensible asset is the distribution a platform can't tax away from you. This week in Behind the CMO: - Why AI retires the product advantage, not the marketing one - The one question that separates an owned moat from a rented cost - How to take this into the 2026 budget conversation Read it free, more in the comments section.
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content