Oho Group Ltd’s Post

What is a computer vision engineer? If you’re interested in #ArtificialIntelligence, #AugmentedReality, and #MachineLearning this topic might be interesting for you.

AI is processing vast amounts of data and simulating new domain of intelligence, offering new perspectives on understanding existence and consciousness 🌟 It has successfully prompted us with to contemplate questions about this nature of reality, and our questionable place in universe 🌎 Whole lot of layers and structures have been opened, giving wide view of Architecture we can have with this vector intelligence, with the expansion architectures will be multiplied with an architecture 👨💼 Do you have what it takes to shape this new verse of future we live in. If you find yourself taking matter in this area to come out in your hands, drop message/comment with your purest original thought on how can your skills and experience help us in generating this new realm 🧠 #hiring #spreadArchitect

3 of 10 #job #openings on 1/11 ✴️ NVIDIA ✴️ Deep learning ✴️ #software engineer & manager #checkitout and #share #opportunity #usa #usjobs #hiring #hiringnow #open #job #reshare #engineeringjobs #engineering #givingback #spreadtheword #connectandgrow

Generative AI Application Architecture - Model references and evaluation models

Optimizing AI/ML Performance: Cutting-Edge GPU Network Designs-Part 3-Parallel and Distributed Computing for the Network Engineer

Retentive Networks (RetNet) is a novel algorithm just published by Microsoft Research revolutionizing the way we deal with transformers. The standout feature? Their time complexity scales linearly with the input sequence, a significant improvement over the quadratic time complexity of traditional transformers. Additionally, the adoption of chunkwise recurrent paradigms enables low-cost inference, which is much more difficult for standard transformers. Microsoft has also released small powerful models like Phi which show the power of high-quality instruction sets. Combined with this new architecture, we could see faster models with better memory efficiency, making LLMs more accessible and paving the way for applications even in resource-constrained environments. #machinelearning #deeplearning #llms #artificialintelligence #generativeai

Large Language Models (LLMs) are rapidly advancing, with daily developments in diverse model architectures, expanding their applications across industries. LLM inference is reshaping data centers, offering higher performance and accuracy, resulting in improved enterprise cost-effectiveness and better customer experiences that drive revenue. However, optimizing LLMs for peak performance requires careful consideration, including factors like parallelism, end-to-end pipelines, and advanced scheduling. It also demands a computing platform capable of mixed precision without compromising accuracy. Enter TensorRT-LLM, an open-source Python API that simplifies the definition, optimization, and execution of LLMs for inference in production. It's your gateway to harnessing the full potential of LLMs in your AI deployments. #AI #LLMs #TensorRTLLM #Inference #AIInnovation

👋 Employee Spotlight 👋   Yu-Hsin Chen is Lead AI Architect at EnCharge AI. He has been working in the field of AI/ML architecture and system design for the past 10 years. Prior to joining EnCharge AI, Yu-Hsin was a Research Scientist at Meta Reality Labs and Nvidia. He received his PhD from MIT in 2018 and his work has also been highlighted in the 50th ISCA Retrospective as one of the top architecture research projects in the past 25 years. Hear from Yu-Hsin on lessons he’s learned from his research, advice he would give to someone entering the field of AI and machine learning, and more.   Q: How did you get started working in computer architecture? A: I have been very intrigued by the inner workings of a computer since I was a kid, which motivated me to study electrical engineering in college and eventually get into the research of computer architectures. As AI and ML started to make rapid advancements in the past 15 years, the idea of designing a computer for the AI of the next decade has become my focus and shaped my career path so far.   Q: What motivated you to take a chance on EnCharge? A: I have been working on cutting-edge AI system design and research for many years. But when I first got introduced to EnCharge, I was truly amazed by the potential of their technology to make an outsized impact. I was also very impressed by the team, which is the most critical factor to success. Overall, I think it is a rare opportunity for me to grow professionally, so I took the plunge.   Q: Who is someone that really contributed to your professional development during your career? A: I am really lucky to have met many great mentors at every step of my career, but my PhD advisors have really contributed to my professional development and shaped who I am today. They have taught me how to deal with unknown and complex problems and gave me an idea of what great leadership looks like. In a way, I’ve decided to join EnCharge as I’m seeking an opportunity to experience the same kind of growth, which I’ve found very rewarding.   Q: For anyone who’s starting a career and looking to get into AI and machine learning, what would you tell that person? A: Always look far and wide and stay curious. As the field of AI keeps moving at a really fast pace, an architect has to have the capability to understand bottlenecks, analyze trade-offs across the whole stack, and be prepared for surprises. Therefore, one needs to know what’s new coming up in the horizon and be able to understand challenges beyond their day-to-day work.   Q: What do you like to do outside of work to relax? A: I enjoy traveling, good food with friends and family, and photography.   Q: What’s something you’d like to accomplish in life that doesn’t have anything to do with work? A: Just be happy with life and be able to spend quality time with the important people in my life. 

Journey Series for Generative AI Application Architecture - Model references and evaluation models