About
Open…
Articles by Xin
-
Coding Mock Interview 专场
Coding Mock Interview 专场
你想旁听一场 真实的 coding interview 过程吗？ 你想知道面试官到底在想什么吗？ 你想弄清楚面试到底是怎么一回事吗？ 你想知道要表现到什么程度才能是 strong hire 吗？…
3 Comments
Activity
16K followers
-
-
-
-
-
Xin Yao shared this中美握手，华尔街欢呼？下周美股或迎来全面反弹 - 中方发布会：中美经贸高层会谈取得了实质性进展，达成了重要共识中美握手，华尔街欢呼？下周美股或迎来全面反弹 - 中方发布会：中美经贸高层会谈取得了实质性进展，达成了重要共识
-
Xin Yao liked thisXin Yao liked thisFrom tools to serve our customers, to tech that improves insights, we're advancing mobility so you advance your business.
-
Xin Yao liked thisXin Yao liked thisWe have a rare opportunity for a director role to lead our LLM platform team. This team powers all LLM needs for Snowflake and our customers, playing a central role in Snowflake's AI journey. We are seeking a candidate with strong technical expertise in inference, along with significant experience in management and strategy. Please take a look, and I appreciate referrals. https://lnkd.in/gjeWRVVA
Experience & Education
-
Flutter Quill
View Xin’s full experience
See their title, tenure and more.
or
Projects
-
Bullet Journal
- Present
See projectAn organizational system that helps you keep track of everything in your busy life
https://bulletjournal.us/home/index.html
BulletJournal is an open source platform for notebook keeping, ledger management, task/project management and coordination with speciality in personal organization, scheduling, reminders, to-do lists, notes sharing, multi-person ledger and team project collaboration. -
Photon Controller
See projectPhoton Controller is an open-source system for managing hardware, containers, and clusters at scale.
Photon Controller is designed to support:
High Scale: Manage tens of thousands of compute nodes in one deployment.
High Churn: Handle thousands of concurrent API requests.
Multiple Tenants: Allocate and manage resources for many users and groups in a single deployment.
Container Frameworks: Easily spin up instances of Kubernetes, Mesos, and Docker Swarm in…Photon Controller is an open-source system for managing hardware, containers, and clusters at scale.
Photon Controller is designed to support:
High Scale: Manage tens of thousands of compute nodes in one deployment.
High Churn: Handle thousands of concurrent API requests.
Multiple Tenants: Allocate and manage resources for many users and groups in a single deployment.
Container Frameworks: Easily spin up instances of Kubernetes, Mesos, and Docker Swarm in seconds.
Languages
-
Chinese
Native or bilingual proficiency
-
English
Full professional proficiency
Other similar profiles
Explore more posts
-
Kaiming Wan
#AutoMQ use cases keep coming in from the community lately — huge thanks to everyone who’s been sharing real-world scenarios and feedback. 🙌 One thing people keep asking me: “Can I just drop in AutoMQ where I’m running Kafka today?” Yes. AutoMQ is 100% #Kafka compatible — APIs, protocols, ecosystem. In most cases, you can swap out your existing Kafka cluster with AutoMQ without changing your application code. What you get on top of that is everything our Diskless architecture brings to the table: ⚙️ 𝐄𝐥𝐚𝐬𝐭𝐢𝐜𝐢𝐭𝐲 𝐛𝐲 𝐝𝐞𝐬𝐢𝐠𝐧 – scale up and down with workload, not with guesswork 📉 𝐌𝐮𝐜𝐡 𝐥𝐨𝐰𝐞𝐫 𝐬𝐭𝐨𝐫𝐚𝐠𝐞 𝐜𝐨𝐬𝐭𝐬 – no more keeping tons of local disks just in case 🔁 𝐙𝐞𝐫𝐨 𝐜𝐫𝐨𝐬𝐬‑𝐀𝐙 𝐭𝐫𝐚𝐟𝐟𝐢𝐜 𝐟𝐞𝐞𝐬 – architecture avoids the classic Kafka replication tax 🧹 𝐒𝐢𝐦𝐩𝐥𝐞𝐫 𝐨𝐩𝐞𝐫𝐚𝐭𝐢𝐨𝐧𝐬 – fewer moving parts, no more wrestling with storage-heavy brokers For teams running Kafka at scale, these differences show up directly in TCO, reliability, and day‑2 operations. Curious to see more real production stories: How are you running Kafka today, and what’s your biggest pain — cost, elasticity, or ops complexity?
-
Steve Love
I've been discussing this recently with Frances Buontempo in the context of coding assistants. We hear a lot from developers saying that LLM assistants in their fave IDEs are great for generating a lot of the (boring) boilerplate code needed to stand up a non-trivial system. However, most popular IDEs have been able to do exactly this for decades--whether with "wizards", code-block and project templates, or other various forms of meta-programming. The difference is that the "old" way is much less hungry of resources, while simultaneously being much more predictable and reliable at automatically generating the "boring" stuff. Coding assistants are also good at other things, of course, but many of the popular (in my conversations with developers) tasks are those I would much prefer to do myself, such as writing tests.
6 Comments -
Fizz Orange
I enabled Bedrock on a Control Tower account locked to us-east-1. Direct model calls worked. Then someone tried us.anthropic.claude-sonnet-4-6 and got AccessDenied. Same account, same role, same region. The us. prefix is a US inference profile — dynamic capacity routing. When you call it in us-east-1, Bedrock tries to serve locally. If capacity is tight, it hops to us-east-2 or us-west-2 over the backbone. No extra cost, no data stored remotely. But Control Tower sees that hop as an API call in a denied region and kills it. I have written before about how a separate Allow SCP can extend Control Tower guardrails. It works for ELB and SES because their API endpoints live in your home region — the SCP evaluator still sees us-east-1. Bedrock inference profiles break that pattern. The cross-region hop is a real API call in us-east-2, and the deny fires on the region condition, not the action. Different evaluation path entirely. I added bedrock:* to the NotAction list in GRREGIONDENY itself. Hit the 5,120 character SCP limit doing it, and had to minify the JSON to make room — json.dumps(policy, separators=(",",":")) strips the cosmetic whitespace after every comma and colon. Six characters saved per two lines, enough across 40+ actions to fit one more service. https://lnkd.in/exZ59JN6 #AWS #Bedrock #PlatformEngineering #ControlTower #IAM
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content