Activity
-
Modern video search systems increasingly rely on multi-vector retrieval to capture visual, audio, and contextual signals. In this technical…
Modern video search systems increasingly rely on multi-vector retrieval to capture visual, audio, and contextual signals. In this technical…
Liked by Qiangjian Xi
-
I’m thrilled to share that I joined 𝗠𝗲𝘁𝗮 𝗦𝘂𝗽𝗲𝗿𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗟𝗮𝗯𝘀 as an AI Research Scientist two weeks ago. I’m excited to…
I’m thrilled to share that I joined 𝗠𝗲𝘁𝗮 𝗦𝘂𝗽𝗲𝗿𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗟𝗮𝗯𝘀 as an AI Research Scientist two weeks ago. I’m excited to…
Liked by Qiangjian Xi
-
What will agentic commerce become in 5 years? And where we all are today? Today we are not limited by AI models. It’s limited by whether our data…
What will agentic commerce become in 5 years? And where we all are today? Today we are not limited by AI models. It’s limited by whether our data…
Liked by Qiangjian Xi
Experience & Education
-
Google
View Qiangjian’s full experience
See their title, tenure and more.
or
Languages
-
Chinese
Native or bilingual proficiency
-
English
-
More activity by Qiangjian
-
Posting our JD -- Senior Staff Data + AI Engineer We are a fast-paced AI startup delivering end-to-end automation and LLM-driven solutions with…
Posting our JD -- Senior Staff Data + AI Engineer We are a fast-paced AI startup delivering end-to-end automation and LLM-driven solutions with…
Liked by Qiangjian Xi
-
From brainstorming to technical docs, Lucid helps engineering teams map out systems and ship code faster. Start diagramming today.
From brainstorming to technical docs, Lucid helps engineering teams map out systems and ship code faster. Start diagramming today.
Liked by Qiangjian Xi
-
[Note: I have since left P-1 AI] Phew, I can finally share what I've been up to since last summer! We just raised a $23 million seed round!! 😅 I…
[Note: I have since left P-1 AI] Phew, I can finally share what I've been up to since last summer! We just raised a $23 million seed round!! 😅 I…
Liked by Qiangjian Xi
-
I'm thrilled to share that I've joined Perplexity AI! I am joining the AI Training team, working on exciting LLM projects! As I wrap up my first…
I'm thrilled to share that I've joined Perplexity AI! I am joining the AI Training team, working on exciting LLM projects! As I wrap up my first…
Liked by Qiangjian Xi
-
Delighted to have graduated from my MS degree in Computer Science from Georgia Institute of Technology with straight As in December. Really enjoyed…
Delighted to have graduated from my MS degree in Computer Science from Georgia Institute of Technology with straight As in December. Really enjoyed…
Liked by Qiangjian Xi
-
In his final book, Henry Kissinger, with former Google CEO Eric Schmidt and former Senior Advisor at Microsoft Craig Mundie, analyzes the rapidly…
In his final book, Henry Kissinger, with former Google CEO Eric Schmidt and former Senior Advisor at Microsoft Craig Mundie, analyzes the rapidly…
Liked by Qiangjian Xi
-
10 predictions for AI in 2025: 1. Massive data is a necessary but not sufficient requirement for superintelligence. It will still be required to…
10 predictions for AI in 2025: 1. Massive data is a necessary but not sufficient requirement for superintelligence. It will still be required to…
Liked by Qiangjian Xi
-
Jeff Bezos on Amazon’s seed round: “It was the hardest thing I’ve ever done.” In 1995, Bezos pitched Amazon to 60 investors, explaining what…
Jeff Bezos on Amazon’s seed round: “It was the hardest thing I’ve ever done.” In 1995, Bezos pitched Amazon to 60 investors, explaining what…
Liked by Qiangjian Xi
-
Attention new grads with exceptional qualification and background, we are opening junior positions and internship roles in coming days. Please check…
Attention new grads with exceptional qualification and background, we are opening junior positions and internship roles in coming days. Please check…
Liked by Qiangjian Xi
-
Join Andrew Ng and Joe Reis today at 9:00 am Pacific Time (in less than an hour!) to discuss how data engineering contributes to data-centric AI and…
Join Andrew Ng and Joe Reis today at 9:00 am Pacific Time (in less than an hour!) to discuss how data engineering contributes to data-centric AI and…
Liked by Qiangjian Xi
-
8 years ago on this date August 8th, we started Nuro in an Airbnb house. Grateful for everyone who supported us and joined us on this journey. The…
8 years ago on this date August 8th, we started Nuro in an Airbnb house. Grateful for everyone who supported us and joined us on this journey. The…
Liked by Qiangjian Xi
-
A moment of celebration and immense change for me and my family. After 25 years of calling the wonderful SF Bay Area home, we’ve relocated to the…
A moment of celebration and immense change for me and my family. After 25 years of calling the wonderful SF Bay Area home, we’ve relocated to the…
Liked by Qiangjian Xi
-
Autoregressive Model Beats Diffusion Llama for Scalable Image Generation paper page: https://buff.ly/3yYkVxK We introduce LlamaGen, a new family…
Autoregressive Model Beats Diffusion Llama for Scalable Image Generation paper page: https://buff.ly/3yYkVxK We introduce LlamaGen, a new family…
Liked by Qiangjian Xi
-
Complex Document RAG with GPT-4o 📑 GPT-4o’s multimodal capabilities means that it’s great at parsing complex PDFs/slide decks with background…
Complex Document RAG with GPT-4o 📑 GPT-4o’s multimodal capabilities means that it’s great at parsing complex PDFs/slide decks with background…
Liked by Qiangjian Xi
-
I'm looking for a senior SWE role in London. Ready to get back at it after enjoying a break last year. Please drop me a line if you know of any…
I'm looking for a senior SWE role in London. Ready to get back at it after enjoying a break last year. Please drop me a line if you know of any…
Liked by Qiangjian Xi
-
If the decoding chip or communication protocol is not unified, Standarzing the port, as USB-C, does not make significant difference. Perhaps, I…
If the decoding chip or communication protocol is not unified, Standarzing the port, as USB-C, does not make significant difference. Perhaps, I…
Liked by Qiangjian Xi
Other similar profiles
Explore more posts
-
Ben (Xiaojun) Li
Google Research Director Denny Zhou, who founded the LLM Reasoning Team at DeepMind, recently gave a great talk at Stanford’s CS25 class: Large Language Model Reasoning. This was one of the most intuitive talks about reasoning model training and application. He outlined new directions for training models to handle questions where answers aren’t easily verifiable. My team is currently working on an enterprise LLM project focused on using reasoning to extract relevant context and generate responses for unverifiable cases. It’s a tough but promising area, and it’s an interesting time to be building. [Link to Denny Zhou's talk: https://lnkd.in/gdJ6Mzi8]
8 Comments -
Jeff Allen
GTC 2026. I spent four days at NVIDIA GTC 2026 — 30,000 developers. My first-hand observation: the inference and orchestration layer is done. Agents, agentic frameworks, inference optimization, orchestration layers — the floor was wall to wall with them. And they all ran on the same models. Peel that back — when every team runs on the same models, the same frameworks, and the same compute infrastructure, the only thing that makes any of them different from one another is the data they use. In 2026, they are all fighting for thin margins in a crowded room. Data is the last remaining moat. I suppose it always was and now we know why. The more commoditized the compute layer becomes, the more true that gets. So given all of that — where were the data platform companies at GTC? And then it hit me — of course they weren't there. Data is the moat. You don't put your moat on an expo floor. The gap is the data pipeline itself. Who builds it, maintains it, and closes the loop back to model performance for the thousands of AI teams that can't do it themselves? NVIDIA told us exactly how real that gap is. Jensen announced the Physical AI Data Factory Blueprint — an open reference architecture for building data pipelines for physical AI. It tells you what to build. And on top of that, every physical AI system Jensen showcased — autonomous vehicles, surgical robotics, humanoid robots, industrial inspection — runs on video. Video is the hardest data type to get right. Temporally complex, legally treacherous, expensive to acquire at quality, and structurally unready when it arrives. Most AI teams don't have the engineering resources to assemble this themselves, maintain it, keep it current, and close the loop back to model performance. And even the giants — Google, Tesla — have built expensive, fragile, proprietary processes with no closed loop back to model performance. The managed platform that makes Jensen's Blueprint operational — at scale, with video at its core — hasn't existed. Until now - that's exactly what we are building. GTC 2026 didn't just validate the problem. It validated the urgency. Physical AI is moving from ambition to commercial deployment — and the data pipeline that feeds it needs to be operational now, not assembled from scratch by every team independently. We're excited about what's ahead. #GTC2026 #PhysicalAI #VideoAI #MLOps #NVIDIA #Flikforge
6 Comments -
Joe Pierce
Most teams running LLM inference are over-provisioning GPUs “just in case.” We benchmarked what happens when you combine NVIDIA NIM with NVIDIA Run:ai with the same inference workloads on dramatically fewer GPUs — with minimal throughput loss. How? ✅ GPU Fractions + Bin Packing: Run multiple NIM models on shared GPUs with true memory isolation — consolidate your deployments, increase density, and reduce your compute footprint. ✅ Dynamic GPU Fractions: Let models burst into unused headroom during traffic spikes — no extra GPUs required. ✅ GPU Memory Swap: Keep infrequently used models warm in CPU memory to slash first-request latency. ✅ Inference Priority Scheduling: Prioritize user-facing inference while training flexes around it. The takeaway: You may not need more GPUs — just smarter orchestration of the ones you have. 🔗 Full benchmarks and methodology in the tech blog: https://bit.ly/3NcIPwL
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content