Kevin Di
San Francisco, California, United States
9K followers
500+ connections
About
Through my years of experience, I have worked in different countries and teamed with…
Activity
9K followers
-
Kevin Di reposted thisKevin Di reposted thisIt's science fiction, minus the fiction! Starting today, Austin riders can be matched with a Waymo autonomous vehicle – only on the Uber app – making their next trip even more magical. With Waymo’s tech and Uber’s proven platform, we can’t wait to introduce our riders to a future of transportation that is increasingly electric and autonomous. Ahead of today’s launch, I took a spin with Good Morning America, where we discussed the Uber-Waymo partnership, the future of transportation, and why 2025 is the year for autonomous tech.Uber expands driverless Waymo cars to Austin, announces plans for more citiesUber expands driverless Waymo cars to Austin, announces plans for more cities
-
Kevin Di posted thisThe Uber Observability team has two openings. We are an agile team composed of individuals with diverse backgrounds, from companies such as Google, Microsoft, Amazon, LinkedIn, Facebook, Cisco, and many startups. We dream big, aim high, and practice grounded execution. Join our team to solve tough large-scale distributed system challenges and improve observability for Uber. (Please apply directly with below links through Uber job portal) Sr Software Engineer on Logging Team https://lnkd.in/df_KBqdr Software Engineer II on Tracing team https://lnkd.in/d2P-p33K
-
Kevin Di shared thisWould love to chat if you are also interested in Observability (logging, tracing, and metrics). https://lnkd.in/ddjM2_V9
-
Kevin Di shared thisKevin Di shared thisKe Jin #Kudos I really appreciate you #GoingAboveAndBeyond in your internship with our team this summer! It’s been great getting to know you, and seeing you finish your project AND fix so many longstanding bugs. Have a great final semester!
-
Kevin Di shared thisKevin Di shared thisCome visit the LinkedIn booth today at #GHC18 to get some free Learning courses, learn about #linkedinlife, and what it’s like to do #productmanagement at LinkedIn. Ask me about the APM program! :) See you there!
-
Kevin Di shared thisShifting Boundaries of IoT with Connected Car DevicesShifting Boundaries of IoT with Connected Car Devices
-
Kevin Di reacted on thisKevin Di reacted on thisFor my chronic illness & cancer/chemo folks 💚 Yesterday: speak at a sold-out conference room Today: resting from post-cancer infusion at hospital (okay, I spoke on Tues, but you get the idea. sending love to anyone else who lives in both worlds) Written without AI from USC Keck hospital in Los Angeles! *** Personal branding is community building with your name on it. I wrote the book on it, literally. ♻️ repost and ✅ follow Goldie Chan. Personal Branding for Introverts (Hachette) is out now. 💚
-
Kevin Di reacted on thisKevin Di reacted on thisToo many apps. Too many decisions. At GO–GET 2026, we shared how we’re taking the next step to simplify everyday life—bringing more of what you need into one place on Uber. Before Uber, I spent over a decade at Expedia Group helping people book travel. Of course, travel isn’t just about booking. It’s everything that happens before, during, and after the trip, and it still feels far too fragmented. Uber is already part of that journey for 200M+ people each month. We’re often the first app people open when they land in a new city. Now we’re expanding that role: helping you move, discover, and get what you need, all in one place. One thing I’m VERY excited about: Hotels are now on Uber. In partnership with Expedia, Uber users in the U.S. will soon be able to search, book, and stay at more than 700,000 hotels globally—right from the app you already use to get around. Even better: Uber One members will earn credits back on every booking, with additional savings across thousands of properties. We’ll also be adding homes through Vrbo later this year. It does feel like a full circle moment for me, and I’m super excited to bring the experience together for our customers in a way that just works. The goal is simple: make complexity feel easy—and give people their time back. Check out our post below to learn more about all of our exciting announcements. https://lnkd.in/gJjW7MSD
-
Kevin Di liked thisKevin Di liked thisThe GPT model divorce is done. Co-parenting begins. The ecosystem wins.🎉 This is a win for Microsoft, a win for OpenAI, and a win for the ecosystem (see exciting news from Gen too in the comment section.). Today’s breaking news is clear: Microsoft and OpenAI renegotiated their partnership to remove exclusivity. OpenAI can now “serve its products across any cloud provider” and has “ended its exclusive licensing agreement” with Microsoft, according to Reuters and Business Insider. This is no longer a marriage. It is co-parenting. I predicted a breakup a couple of years ago here on Linkedin. :) The first phase required two parents. Compute and research. Microsoft brought hyperscale infrastructure. OpenAI pushed frontier models. Together, they produced GPT. That phase is over. They still share the existing stack. Microsoft keeps access through 2032 and remains deeply integrated. But both are now free to build their own future. Separate roadmaps. Separate bets. You can already see it. Microsoft is expanding its ecosystem. Its Copilot platform is integrating with Engine by Gen and delivering compliant financial intelligence directly into AI workflows. The writing is on the wall that companies like Gen are moving closer with frontier labs. Tighter collaboration with model builders. Deeper integration at the research layer. More direct alignment with where models are going. We are welcoming our first frontier model researcher to my team too!! Last but not least, Divorce is never easy. But in this case, it is exactly what the market needed. Go OpenAI, Microsoft, and Gen! Joe Talbott Charlie Lin Ben Wadors Vincent Pilette Leyla Bilge xAI Anthropic Google
-
Kevin Di liked thisKevin Di liked thisCLP 0.11.0 is here. This release adds support for more flexible AWS credential configuration in the Helm chart and support for deploying Presto with the CLP connector. Additionally, the log ingestion pipeline has been optimized for better real-time performance, alongside various bug fixes and general enhancements. Thanks to Naman Goyal (Amazon), Qingqin Hou and Kevin Di (Uber), Xu Zhao (Meta), and other members of the community for their deployment feedback, feature requests, and bug reports. https://lnkd.in/gUpFDynp
-
Kevin Di reacted on thisKevin Di reacted on thisWe're excited to release CLP v0.10.0. Users can now query multiple datasets simultaneously from both the UI and the CLI. We've also built strong fault tolerance into our log-ingestor for greater reliability. CLP's Presto connector now supports custom S3 endpoints, allowing integration with S3-compatible storage providers, such as MinIO. It also includes other improvements and bug fixes. Thanks to Naman Goyal (Amazon), Qingqin Hou and Kevin Di (Uber), Xu Zhao (Meta), Theo Sherman (OneAxiom), and Sudheer Gajula (Index Exchange) for their deployment feedback, feature requests, and bug reports. https://lnkd.in/gd3_RMy6
-
Kevin Di reacted on thisKevin Di reacted on thisMost people misunderstand what “AI-native” means. They think it’s a new stack. A new set of tools. A new workflow template. It’s not. AI-native is a pricing shift—and once you see it, your habits have to change. The core idea In the old world, code was expensive. Not because files are expensive, but because producing code required scarce human time: hiring engineers, coordinating teams, debugging, maintaining, communicating. So we built a whole professional culture around conserving code: - maximize reuse - avoid one-off scripts - don’t build “throwaway” tools - rely on senior intuition when visibility is costly In the AI world, that assumption breaks. Code becomes cheap. Not perfect. Not magical. But cheap enough that the optimal strategy flips. The analogy that makes it stick Aluminum used to be more valuable than gold—not because it was rarer, but because extracting pure aluminum was costly. Then production changed (electrolysis + scalable refining). Aluminum became cheap. Nothing about Earth’s geology changed. Only the cost of production changed. AI is doing the same thing to code. What changes when code is cheap? When “making a tool” is cheap, you stop treating code like gold. You start using code like aluminum: - generate quick visualizations instead of “debugging by intuition” - create disposable dashboards and mini-apps to surface evidence - instrument everything because the marginal cost is low - turn vague problems into observable problems Here’s the key line: Truth is gold. Code is aluminum. Truth remains expensive. Code becomes the inexpensive lever you spend to buy certainty. The practical habit shift Old habit: “Don’t write code unless it’s reusable.” AI-native habit: “Write code to replace time and buy certainty—even if it’s single-use.” This is not anti-engineering. It’s economics. If a “throwaway” tool takes 10 minutes and saves half a day of guessing, it’s a great trade. I wrote an article that unpacks this with a concrete story: building “burn-after-reading” software to inspect messy data—faster, more precise, and more reliable than intuition-driven debugging. If you’re building with AI and still working like code is expensive, this will feel like a reset.
-
Kevin Di liked thisKevin Di liked thisBelief Builds Strength, Not Skill. True leaders don't just share tactics; they ignite confidence. Before the training, before the strategy… They make you believe in yourself. I remember a pivotal moment early in my career. I was in a room where I felt out of place—everyone else seemed sharper, more experienced. I was silent, fearing I’d be found out. Then, a leader turned to me and said, “I want your perspective. Say it like you mean it.” No grand speech, just unwavering belief placed on my shoulders. And at that moment, something shifted. When someone sees your potential before there’s proof, you start moving like success is inevitable. That’s why the "less qualified" often surpass the "more talented." They were given permission to fully commit. Skills can be taught. Confidence? It’s contagious. Be the leader who instills belief, and watch others achieve the impossible. For more posts like this, make sure you FOLLOW Emily Pappas!
-
Kevin Di liked thisKevin Di liked thisBack in 2022, I joined Uber with the mission to help build a secure service mesh from the ground up. It’s been quite a journey — full of engineering challenges, design decisions, and lessons learned. This week, I’m excited to share this journey together with Matt Mathew at KubeCon NA — the architecture, challenges, and real world lessons we’ve learned along the way. Join us if you’re interested in developing and adopting large-scale service mesh and zero-trust authN/Z solution in production! https://lnkd.in/gCCey7ha #kubecon #envoy #spire #servicemesh #authentication #authorization #zerotrust #accesscontrol #securityKubeCon + CloudNativeCon North America 2025: Authenticating and Authorizing Every Con...KubeCon + CloudNativeCon North America 2025: Authenticating and Authorizing Every Con...
Experience
-
-
-
-
-
-
-
-
-
Education
-
-
-
-
Licenses & Certifications
Volunteer Experience
-
Technology Trainer
Community Technology Network
- 4 months
Education
Technology trainer at AmericaWorks. Teaching people how to use modern technologies including smartphones, computers, and software.
Courses
-
2D and 3D Animation
10
-
3D Modelling and Reconstruction
10
-
Advanced Computer Graphics
10
-
Introduction to Games Design
10
-
Multimedia Information Management
10
-
Virtual Reality
10
Honors & Awards
-
Dean's List Equivalent Scholarship
Heilongjiang University
Languages
-
English
-
-
Chinese (Simplified)
-
Organizations
-
Toastmasters Club of Singapore
-
- PresentActively participate in Toastmasters social events and give public speeches to develop presentation and communication skills.
Other similar profiles
Explore more posts
-
Abhiyan Khanal
I use Claude Code to ship my engineering tasks in weekdays, but its potential goes far beyond just refactoring or debugging code. I recently built a workflow where I can publish a complete song to YouTube using Claude Code from the CLI. The entire pipeline, lyrics composition, local AI music generation, cover art design, video assembly, and YouTube publishing, runs from a terminal with a prompt and minimal human interaction. What used to require coordinating across multiple tools, multiple skill sets, and multiple hours now takes minutes from a single interface. But here's the bigger picture: this pattern applies everywhere. The same approach of chaining AI capabilities into end to end workflows is transforming content creation, marketing, research, design, and operations across industries. We're entering an era where the competitive advantage isn't just knowing how to use individual tools. It's knowing how to orchestrate them into workflows that collapse entire pipelines into simple interactions. What workflows in your industry are ready for this kind of transformation? https://lnkd.in/gN_tR6tz PS: This post was generated using AI, and the quality of song could be better using models with more parameters, and better prompting maybe? but I used distilled model due to resource constrains. #AI #FutureOfWork #Automation #ClaudeCode #MusicGeneration #Innovation #Productivity
-
Manas Jade
🚘💡 Edge AI is reshaping industries—and redefining the talent we need to build the future. Our EVP & President of Software, Advanced Safety, and User Experience, Javed Khan, shares how Edge AI is transforming real-time decision-making across autonomous vehicles, aerospace systems, robotics, and beyond. 🔗 Read the article here: https://bit.ly/40AM3hc #ImpactEveryJourney #FutureofMobility #Innovation #EdgeAI
-
Conduktor
Who needs self-service? In this article, Conduktor’s VP of Engineering, Matt Searle, talks about his own experiences attempting to balance developer autonomy alongside platform stability. His key insight? Govern Kafka, or risk havoc. Read his latest blog—and learn more about how Conduktor can help you implement self-service with guardrails. https://lnkd.in/ejWTRctu
-
Kalyan Ravilla
CocoaPods is not deprecated, but Swift Package Manager (SPM) is clearly the future. While CocoaPods continues to work and is still supported, Apple has been consistently encouraging developers to adopt Swift Package Manager as the preferred dependency management solution. SPM is deeply integrated with Xcode, requires no additional tooling, and offers a simpler and more native workflow for managing dependencies. As a result, many modern iOS and cross-platform projects are gradually migrating away from CocoaPods. CocoaPods remains relevant for legacy projects, but for new development, SPM is the recommended path forward. #iOS #SPM #COCOAPODS
-
Sunil Mamilla
As LLMs evolve from static models into dynamic components of real-world applications, they must do more than perform tasks; they must also adapt to context, personalize interactions, and continuously improve through feedback. This blog outlines a practical, modular approach to designing adaptive, production-ready LLM systems using: - Supervised Fine-Tuning (SFT) for core task competence - Reinforcement Fine-Tuning (RFT) for behavioral alignment - Context-aware routing for persona and adapter control - RAG for real-time knowledge injection - Telemetry loops for continuous learning and refinement It also includes a curated toolchain (LoRA, trl, vLLM, Langfuse, and more). https://lnkd.in/gBtjeg_X
2 Comments -
Bilal Bashir
Spoke with an engineer in a 1:1 today. He shared an idea to improve our team's velocity. He proposed that since we had multiple services we should split the team for each microservice to make people more productive. I explained to him we don't have microservices but we do have a micromanager so there is no need to split the team. #leadership 🤗
2 Comments -
Darlington Moyo
🧩 Solving GraphQL Pagination in Matillion Custom Connector using Script Pagination Read it here: https://lnkd.in/gVTRBiEb GraphQL APIs are elegant, until you have to deal with pagination. In this article, I break down how we approached cursor-based pagination inside the Matillion Custom Connector framework, using script pagination to handle complex API responses and dynamic cursors. It covers: - The difference between GraphQL and REST pagination models - How to pass dynamic variables through queries - An example using a real API - How the approach scales across connectors in Matillion's Data Productivity Cloud If you've ever designed or debugged GraphQL integrations, you'll probably relate. Would love to hear how you've handled pagination challenges in your own pipelines 👇 #Matillion #GraphQL #DataEngineering #DataIntegration #CustomConnectors #ETL #API #Pagination #DataProductivity
-
Jenom S.
Building a Production-Ready Two-Tier Architecture on AWS. I’m excited to share my latest article where I delve into designing a robust, scalable, and secure two-tier architecture on AWS for Adventure Tour Inc. This implementation aligns with industry best practices, ensuring high availability, seamless scalability, and stringent security measures. Key Highlights: Auto Scaling Infrastructure: I utilized Auto Scaling Groups behind an Application Load Balancer to ensure elasticity and high availability. Private Database Tier: Deployed Amazon RDS MySQL within private subnets, enhancing data security and isolation. Security Best Practices: Implemented multi-layered security groups, IAM roles, hardened Golden Image and Systems Manager for secure access management. Comprehensive Monitoring: Integrated Amazon CloudWatch for detailed monitoring, enabling proactive scaling decisions. This architecture not only supports the current needs of Adventure Tour Inc. but also provides a scalable foundation for future growth. Whether you’re looking to enhance your AWS infrastructure or seeking insights into best practices, this guide offers valuable perspectives. Read the full article on my medium link below and let me know your thoughts. Broadus Palmer Joe Stephens Janita Green Williamson Brett D. Tait Hoglund Troy Ingram #AWS #CloudArchitecture #DevOps #Security #Scalability #InfrastructureAsCode #CloudComputing #AWSArchitecture
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content