Stanford Online

🎉 Excited to Share a Milestone 🎉 After two years of learning and exploration, I am honored to share that I have successfully earned the Stanford Artificial Intelligence Graduate Certificate! Over the past two years, while pursuing my PhD in Architecture at The University of Auckland, I completed four graduate AI courses on Stanford Online, including: • CS229: Machine Learning (Taught by Chris Ré and Tengyu Ma) • CS224N: Natural Language Processing with Deep Learning (Taught by Yang Diyi and Tatsunori Hashimoto) • CS224W: Machine Learning with Graphs (Taught by Jure Leskovec) • CS231N: Deep Learning for Computer Vision (Taught by Ehsan Adeli and Fei-Fei Li) A heartfelt thank you to these exceptional professors for delivering such insightful, engaging, and inspiring courses. Your lectures not only deepened my understanding of artificial intelligence but also sparked ideas that I will carry into my future research. My journey into AI began in 2021 when I first encountered Andrew Ng's Machine Learning course on Coursera, which opened my eyes to an entirely new world of possibilities. For someone from an architectural background like myself, that course was a revelation, opening the door to the fascinating world of artificial intelligence. Thank you, Andrew Ng, for inspiring me to embark on this path and showing how AI can revolutionize fields far beyond computer science. This accomplishment would not have been possible without the support of my PhD supervisor, Manfredo Manfredni, from the School of Architecture and Planning, and the invaluable guidance from my doctoral thesis advisor, Kaiqi Zhao, from the School of Computer Science at The University of Auckland. Thank you for helping me bridge the worlds of architecture and artificial intelligence in my doctoral research. Looking ahead, I believe that the next breakthroughs in architecture may lie in the integration of novel AI technologies with architectural theories and design approaches. This belief drives my research as I strive to explore meaningful and innovative ways to combine AI with the study of architecture, cities, and spatial environments. 🚀 #AI #Architecture #StanfordUniversity #ArtificialIntelligence #MachineLearning #DeepLearning #PhDLife #Innovation #ArchitecturalInnovation #ContinuousLearning #UoA #ComputerScience

What a nice birthday present! My research paper "ELECTRA and GPT-4o: Cost-Effective Partners for Sentiment Analysis" was just published on arXiv. https://lnkd.in/gCX7gZhN This all started as an interesting observation during my first assignment at XCS224U. I had just finished fine-tuning ELECTRA on sentiment, but I wasn't happy with the performance. I tried prompting GPT-4o-mini via API, and it was about the same. We were learning DSPy around this time, and I was curious: what if I shared the ELECTRA prediction with GPT-4o-mini, would it help? Turns out it did. Augmenting GPT-4o-mini prompts with ELECTRA predictions significantly improved performance over either model alone, and with the lowest cost/performance ratio. Later, I found that a fine-tuned GPT-4o performed the best, but a fine-tuned GPT-4o-mini was nearly as good -- and at a quarter of the cost.

I’m happy to share that I’ve obtained a Professional Certificate in Artificial Intelligence from Stanford University through Stanford Online! This fall I had the rare opportunity to take a break from work and make a career pivot by working toward this new credential through Stanford. The courses are challenging, but extremely interesting and well structured. Each course took me at least 15-20 hours per week for 10 weeks. I do NOT recommend taking three courses at the same time like I did unless you are able to make a significant time commitment. Special thanks to François Chesnay for his invaluable office hour sessions and Slack support, and to Christina Zuno and Daud Gill for their exceptional program management. These courses have equipped me with the skills needed to venture into the field of AI professionally. Excited for the opportunities ahead in my career! Courses completed: - XCS221 - Artificial Intelligence: Principles and Techniques - XCS224W - Machine Learning with Graphs - XCS224N - Natural Language Processing with Deep Learning #ArtificialIntelligence #ProfessionalDevelopment

🌟 Finished: Business Opportunities and Applications of Generative AI – Stanford Online! 🚀 Hi Community, Excited to share that I’ve completed Stanford Online’s XFM111: Business Opportunities and Applications of Generative AI. If you’re curious about how generative AI can drive real-world impact, this course is an exceptional deep dive into applying AI in business settings while addressing its opportunities and challenges. Here are two standout insights I took away: 📌 Agentic AI Workflows – AI systems that proactively assist by taking on tasks autonomously while maintaining human oversight. A glimpse into the future of collaborative AI. 📌 The Turing Trap – A thought-provoking framework highlighting the risks of over-automation and advocating for augmenting human capabilities instead. As I think about applying these ideas to my projects like Milestones Insights (milestonesinsights.com) and my work at Boeing, this course has sharpened my understanding of how AI can be designed not just to replace tasks but to empower individuals and teams. 💡 Why this matters: Generative AI is a tool with immense potential—but to unlock it, we need to understand how to deploy it ethically, responsibly, and effectively. This course provides actionable strategies for integrating AI into organizational workflows, fostering trust, and navigating legal considerations. If you’re interested in how AI can transform industries or want to explore the human-centric side of AI, I’d be happy to share more. Big thanks to Stanford Online and the incredible instructors for such a practical and thought-provoking learning experience. Here’s to designing AI systems that not only innovate but also inspire. After all, The Effort is the Reward. #StanfordOnline #GenerativeAI #AgenticAI #TuringTrap #FutureOfWork #Innovation #ResponsibleAI