About
Seasoned engineering leader with experience leading ML and fullstack teams to achieve…
Activity
9K followers
-
Di Li shared thisThis is what it looks like when a company means what it says. Anthropic has been clear from the start: we support lawful national security uses of AI, but we won’t enable mass surveillance of Americans or fully autonomous weapons. Today’s statement makes me even more proud to be part of this team. https://lnkd.in/gA79r9k2Statement on the comments from Secretary of War Pete HegsethStatement on the comments from Secretary of War Pete Hegseth
-
Di Li shared thisHiring - hit me or Marisa Jones if you see a fit here. https://lnkd.in/gZgf3qPzStaff Backend Engineer, Merchandising Platform - Careers at AirbnbStaff Backend Engineer, Merchandising Platform - Careers at Airbnb
-
Di Li shared thisTime to explore more new offerings at AirbnbDi Li shared thisNow you can Airbnb more than an Airbnb. Airbnb Services: Book the world’s best chefs, trainers, massage therapists, and more Airbnb Experiences: Completely reimagined, and hosted by the locals who know their city best An all-new app, with homes, experiences, and services all in one place Learn more: airbnb.com/release
-
Di Li shared thisHiring Backend and Machine Learning Engineers https://lnkd.in/gPUt7TBV https://lnkd.in/gUDaaMNg Reach out to Erica Cortes Walker or me if interested. #hiring #backend #MLESenior Machine Learning Engineer, Guest & Host - Careers at AirbnbSenior Machine Learning Engineer, Guest & Host - Careers at Airbnb
-
Di Li shared thisExcited to share another team's work around AI-powered use cases boosting computer vision accuracy and performance at Airbnb.Di Li shared thisExcited to share that our latest blog post, Airbnb’s AI-powered photo tour using Vision Transformer, where we show our journey of using Vision Transformers to power our photo tour product, by incorporating pretraining, multi-task learning, ensemble learning, and knowledge distillation: https://lnkd.in/gJi_p5qW Please feel free to reach out to me, Xiaoxin (Aaron) Yin or Di Li if you are interested in our work.Airbnb’s AI-powered photo tour using Vision TransformerAirbnb’s AI-powered photo tour using Vision Transformer
-
Di Li shared thisHiring post - If you are deeply passionate about building world class data/ML driven intelligence platforms that power high quality listings and help host better merchandising them to guests, please apply or reach out to me or Bennett Bontemps Suzy LambertStaff Software Engineer, Data Engineering - Careers at AirbnbStaff Software Engineer, Data Engineering - Careers at Airbnb
-
Di Li shared thisDi Li shared thisWhat is the difference between 𝗟𝗮𝗺𝗯𝗱𝗮 𝗮𝗻𝗱 𝗞𝗮𝗽𝗽𝗮 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲𝘀? Lambda and Kappa are both Data architectures proposed to solve movement of large amounts of data for reliable Online access. The most popular architecture has been and continues to be Lambda. However, with Stream Processing becoming more accessible you will be hearing a lot more of Kappa in the near future. Let’s see how they are different. 𝗟𝗮𝗺𝗯𝗱𝗮. ➡️ Ingestion layer is responsible for collecting the raw data and duplicating it for further Real Time and Batch processing separately. ➡️ Consists of 3 additional main layers: 👉 Speed or Stream - Raw Data here is coming in Real Time and is processed by a Stream Processing Framework (e.g. Flink) then passed to the Serving layer to create Real Time Views for low latency near to Real Time Data access. 👉 Batch - Batch ETL Jobs with batch processing Frameworks (e.g. Spark) are run against raw Data to create reliable Batch Views for Offline Historical Data access. 👉 Serving - this is where the processed Data is exposed to the end user. Latest Real Time Data can be accessed from Real Time Views or combined with Batch Views for full history. Historical Data can be accessed from Batch Views. ❗️ Processing code is duplicated for different technologies in Batch and Speed Layers causing logic divergence. ❗️ Compute resources are duplicated. ❗️ Need to manage two Infrastructures. ✅ Distributed Batch Storage is reliable and scalable, even if the System crashes it is easily recoverable without errors. 𝗞𝗮𝗽𝗽𝗮. ➡️ Treats both Batch and Real Time Workloads as a Stream Processing problem. ➡️ Uses Speed Layer only to prepare data for Real Time and Batch Access. ➡️ Consists of only 2 main layers: 👉 Speed or Stream - similar to Lambda but (optionally) often contains Tiered Storage which means that all of Data coming into the system is stored indefinitely in different Storage Layers. E.g. S3 or GCS for historical data and on disk log for hot data. 👉 Serving - same as Lambda but all transformations that are performed in Speed Layer are never duplicated in Batch Layer. ❗️ Some transformations are hard to perform in Speed Layer (e.g. complex joins) and are eventually pushed to Batch storage for implementation. ❗️ Requires strong skills in Stream Processing. ✅ Data is processed once with a single Stream Processing Engine. ✅ Only need to manage single set of Infrastructure. Have you dealt with Kappa architecture in your day-to-day? What are your thoughts around it? Let me know in the comments 👇 -------- Follow me to upskill in #MLOps, #MachineLearning, #DataEngineering, #DataScience and overall #Data space. 𝗗𝗼𝗻’𝘁 𝗳𝗼𝗿𝗴𝗲𝘁 𝘁𝗼 𝗹𝗶𝗸𝗲 👍, 𝘀𝗵𝗮𝗿𝗲 𝗮𝗻𝗱 𝗰𝗼𝗺𝗺𝗲𝗻𝘁! Join a growing community of Data Professionals by subscribing to my 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: https://lnkd.in/e5d3GuJe
-
Di Li shared thisDi Li shared thisSimply the best high-level talk on LLM so far.
-
Di Li liked thisDi Li liked thisClaude Fable 5 is available today! It's a new moment for AI: a Mythos-class model, the most capable class of systems we've built, now safe for general use. It's already changed how we work internally, and I'm excited to see what you all do with it. Every request runs past safety classifiers trained to detect misuse in cybersecurity and biology. When one triggers, your request is answered by Opus 4.8 instead. More than 95% of sessions never see a fallback, and 1,000+ hours of external red-teaming produced no universal jailbreak. In terms of benchmarks, Fable 5 reached 80.3 on SWE-bench Pro (Opus 4.8 scores 69.2), 88 on Terminal-Bench 2.1. State-of-the-art on nearly every coding benchmark we tested. But the benchmarks undersell how truly capable it is. Fable holds quality deep into long, hard problems where most models degrade. It verifies its own work. It catches what others miss, things like root-cause bugs that no other model had surfaced. Base44 found it "much deeper and better at one-shotting full apps"; at Genspark it came out #1, winning head-to-head against every model they tested. Internally, writing code stopped being the slow part a while ago — Anthropic engineers on average shipped 8x as much code per quarter as they did compared to 2021-2025 — Fable pushes the bottleneck further toward verification and review. We're excited to make all of that available today for every use case outside bio and cyber. For API customers, here's how we've imagined fallbacks: pass a fallbacks parameter and the Messages API retries any blocked turn on Opus 4.8 server-side — even mid-stream, keeping the partial output. We think of this as a graceful handoff between models, and we'll be iterating on the design with the community. Moments like this are worth doing right. We're making sure it's safe, but the classifiers may be annoying at times. They're tuned conservatively, and false positives will keep coming down. Read more here: https://lnkd.in/giBEAAcP
-
Di Li liked thisDi Li liked thisScaling your SaaS shouldn't mean rebuilding from scratch. AWS gives you the cloud infrastructure, AI tools, and go-to-market programs to grow faster.AWS for Software and Technology – ISV Cloud Solutions - AWSAWS for Software and Technology – ISV Cloud Solutions - AWS
-
Di Li liked thisDi Li liked thisAnthropic has confidentially submitted a draft S-1 registration statement to the Securities and Exchange Commission. Pending completion of SEC review, this gives us the option to pursue an initial public offering. Read more: https://lnkd.in/en4aKgVxAnthropic confidentially submits draft S-1 to the SECAnthropic confidentially submits draft S-1 to the SEC
-
Di Li liked thisDi Li liked thisAnthropic has raised $65 billion in Series H funding at a $965 billion post-money valuation, led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital. This Series H financing also includes many of the world's most prominent investors, as well as strategic infrastructure partners and hyperscalers. This investment reflects the continued rapid growth in demand for Claude across enterprises and the people who use it for their everyday work. We will utilize this capital to advance our safety and interpretability research, continue to expand our compute footprint, and scale the products and partnerships our customers rely on. I am grateful to our investors, our customers, and our employees for their continued collaboration and support. Read more: https://lnkd.in/gxmPJfueAnthropic raises $65B in Series H funding at $965B post-money valuationAnthropic raises $65B in Series H funding at $965B post-money valuation
-
Di Li liked thisAfter an incredible 8+ years of leading Product and Design and helping build this world-class company into what it is today, I have decided to leave DoorDash. This will be the last summer with a kid at home full-time, and I know it’s time I want to spend with my daughter before she heads off to college. My last day will be May 22nd, and what comes after that will include downtime, tinkering, and planning for my next play. My journey at DoorDash was shaped by many moments: Building a product customers use every day: DoorDash's customer-obsession culture is world-class. From diagnosing “disaster deliveries” to spending time each month working at merchant counters and making deliveries, the product we built learned from every interaction and stayed focused on one goal - solving our customers' problems. Assembling a team that hustles and delivers: Culture is shaped by the people you hire. Early on, we created a value of never compromising on our bar or the principles that we wanted to uphold with the builders we bring on-board. I am incredibly proud of everyone I got to build with, learn from, and ship meaningful work alongside. Shipping impactful things: Keeping DoorDash operating for our users during the pandemic, building one of largest subscription networks, innovating new ways to work with local businesses, launching the market-leading grocery service, creating a category-defining advertising product, and expanding globally, to name a few. I want to thank Tony Xu for trusting me to be part of this journey. His optimism and high expectations have shaped me forever. Thank you as well to Ryan Sokol, Prabir Adarkar, Ravi Inukonda, Mariana G., Keith Yandell, Tia Sherringham, Elizabeth Jarvis-Shean, Christopher Payne, and Miki Kuusi for being incredible teammates. A final shoutout to the amazing Product and Design teams we built from scratch into what I can confidently say are among the best in the industry. I am privileged and honored to have worked with you all. While I will remain connected as an advisor at the company, to each and every Doordasher, I am so proud of what we have accomplished and built and will be a proud cheerleader forever.
-
Di Li liked thisDi Li liked thisIntroducing Claude Design by Anthropic Labs: a new way to make designs, prototypes, slides, and one-pagers by talking to Claude. Claude Design is powered by Claude Opus 4.7, our most capable vision model. Describe what you want, and Claude builds the first version. Refine through conversation, inline comments, direct edits, or custom sliders, then export to Canva, as PDF or PPTX, or hand off to Claude Code. Claude reads your codebase and design files to build your team's design system, then applies it automatically, keeping every project on-brand. Claude Design is available in research preview on the Pro, Max, Team, and Enterprise plans, rolling out throughout the day. Try Claude Design: claude.ai/design Read more: https://lnkd.in/e3YaGiiA
-
Di Li liked thisDi Li liked thisWhat does an "Elite Founder" look like? Well, according to a16z speedrun's website, you're looking at him! Jokes and randomly-selected-B-roll-photos aside, it has been an incredible 12 weeks taking Amdahl to the next level alongside a community of exceptionally sharp and motivated founders. I've been especially impressed by the support we've had from the a16z speedrun team - their hands on help in hiring, sales, fundraising, and navigating the hard decisions that come at this stage have been an unfair advantage for our small (but mighty!) team - huge thanks to Fareed Mosavat Lejla Johnsen Jonathan Lai Andrew Chen Tom Hammer Macy Mills Bella Nazzari Jordan Carver Jordan Mazer Jacqueline Young Andrew Lee and everyone else who helped us along the way. It still feels like we're just getting started - excited to see where we're at in another 12 weeks!
-
Di Li liked thisDi Li liked thisStoked to share that our run-rate revenue has crossed $30 billion, up from $9 billion at the start of the year! We've also just signed an agreement with Google and Broadcom for multiple gigawatts of compute, coming online starting in 2027. Excited to continue serving the wild growth in demand we're seeing. More about our deal here: https://lnkd.in/ehbz23j6 Separately on the growth team, we're hiring across the board to help Anthropic stay on the exponential. If you're an engineer, PM, designer, or data scientist with a background working on growth teams, get in touch through our jobs page. For growth PMs, right now I'm particularly interested in folks who have a background in either monetization or growing API products.Anthropic expands partnership with Google and Broadcom for multiple gigawatts of next-generation computeAnthropic expands partnership with Google and Broadcom for multiple gigawatts of next-generation compute
Experience
-
-
-
-
-
-
-
-
-
-
-
-
-
Education
Languages
-
English
Professional working proficiency
-
Chinese
Native or bilingual proficiency
Other similar profiles
Explore more posts
-
Chris Balestras
after working with hundreds of candidates at VibeScaling trying to place them at the top AI-native startups in SF/NYC, here are 5 unconventional tactics that we've seen work to stand out 👇🏻: 1️⃣ customer interviews tactic: when referrals fail, go directly to their customers - outbound 30-50 asking about pain points, turn 3-5 responses into a short analysis, and send it to the head of sales. Shows you can create value before you're hired while most people just send resumes and pray 2️⃣ the "not now" drip campaign tactic: when you get rejected for a specific reason, don't walk away - fix the objection and stay in touch. If they say you're "not technical enough," learn their space, use their product, and send updates on your progress. They often come back when timing changes, and you'll be top of mind as someone who actually takes feedback 3️⃣ the brand book tactic: hiring managers have tunnel vision and only see your recent experience. Create a portfolio with career highlights, peer quotes, outbound examples, and deal case studies to tell your full story. Helps you control the narrative and show skills beyond what's on your resume 4️⃣ proactive references tactic: instead of waiting for companies to backchannel you, have a strong reference reach out to hiring managers before your final rounds. Lets someone handle objections for you while most candidates just wait around. Being proactive with references gives you a huge edge 5️⃣the gratitude loop tactic: send thank you notes to everyone who helped - interviewers, referrers, advisors - regardless of outcome. People remember genuine appreciation, and it's a small tech world where today's "no" often becomes tomorrow's opportunity invest in relationships and stand out by doing what others won't wrote the full breakdown with templates in GTMBA 👇🏻
7 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content