Of Passion and Chance: Taking GenAI from Silicon Valley to Sofia
The global interest level for AI is off the chart, especially generative AI (GenAI). So April is going to be a very busy month for me, since I will be traveling pretty much the entire month. I am really excited about my trip to Sofia, as it will be the first time I visit Bulgaria. But the genesis of this trip, however, was anything but planned. It unfolded through a series of chance serendipity, and here is the inside scoop.
Way back in 2021, during the COVID pandemic, I had the opportunity to teach an experimental online AI/ML seminar series for Engineers with Prof. Alex Iliev through the UC Berkeley Extension Program. Since I never liked the administrative part of teaching, Prof Iliev took the role of Lead Instructor for this course where he manages all the work with the university and other staff (including other instructors, the video recording team, and the office of the registrar, etc.). I took the easier job of simply teaching the module on Industrial AI Adoption Maturity where we discussed the challenges throughout this AI adoption journey. Some of the challenges we discussed include, but are not limited to AI trust, AI bias, transparency and explainability, and the greater societal and environmental impacts of AI.
Despite our efforts to promote this online seminar series with blog post and panel discussions (above is a video recording of the panel event), there were some adoption challenges. But this wasn’t a surprise, because many people are still unfamiliar and uncomfortable with virtual instructions through video streaming platforms (e.g. ZOOM, or Microsoft TEAM). Besides, the disruption of the pandemic was just too much for people to handle, and the audience’s attention span was just not there for a series of long lectures.
Despite the course’s modest success, this teaching venture forged a connection between Prof. Iliev and me, rooted in mutual respect and appreciation for our respective strengths and passion. Besides teaching at UC Berkeley Extension, Alex is also a Professor and Academic Head at SRH Berlin University of Applied Sciences, and he’s also a member of the faculty at the Bulgarian Academy of Science (BAS), a highly reputable research institution in Bulgaria. Coincidentally, PROS also has an engineering office in Sofia through our acquisition of Vayant Travel Technologies in 2017. Naturally, we explored the possibility of having me as a guest lecturer at SRH Berlin Univ. and BAS. However, the lingering pandemic thwarted our plans.
Guest Lecture at the Bulgarian Academy of Science – Institute of Mathematics and Informatics
Today, 3 years later (2024), with COVID firmly behind us, the opportunity to realize these plans resurfaced as we catch up. And this time, we are going to make it happen! This week, PROS and the Institute of Mathematics and Informatics (IMI) at BAS have just signed a memorandum that kicks off our collaboration. And I am thrilled to be giving 2 guest lectures at IMI/BAS at our inaugural event: Decoding GenAI (If you are interested, you should REGISTER NOW, as space is limited). Not to mention that I am also eager to meet my fellow colleagues in Sofia, many of whom I’ve never met before.
The first lecture will be a general lecture for anyone interested in having a deeper look into GenAI (e.g. ChatGPT or any large language model). Since BAS is a research institution, I think it’s appropriate that we delve into the science behind the seemingly magical power of GenAI. Not only will we understand why hallucination occurs, we will show you how to keep it under control. So you can experiment and leverage GenAI safely and productively in your organization.
Title: Decoding GenAI: The Science Behind the Magic of ChatGPT
Abstract:
ChatGPT is akin to modern sorcery. With it, no subject is beyond the reach of the curious, including Generative AI (GenAI). However, ChatGPT often fails to offer an intuitive yet deep enough understanding of GenAI that enable leaders to leverage it effectively in business. In this captivating learning experience, we aim to shed light on the science that powers GenAI. Rather than telling you that GPT is a large language model (LLM), we will show you how LLM works, how it’s trained, what’s the training data, and why they will all hallucinate. Despite the certainty of hallucination, we will also provide a technique that enables businesses to use ChatGPT safely.
Moreover, we will go beyond ChatGPT and examine why GenAI is able to create captivating dialogues, stunning artwork, and melodious tunes that ignite the imagination. Throughout this exploration, we will also navigate the diverse landscape of GenAI’s business applications, including some great and not-so-great use cases. Through this interactive learning experience, we will “almost” make you an expert in GenAI by equipping you with the knowledge needed to wield this transformative technology responsibly to drive business innovation and growth.
Although GenAI exhibits truly extraordinary abilities (many of which we don’t quite understand completely), there are known limitations that restrict GenAI’s applicability in business, especially in mission-critical decisions. The second lecture is devoted to exploring these limitations and how we can use GenAI with other AI technologies to fill this gap. This is crucially important for business, as it will turn GenAI from merely a tool for productivity into an instrument for profitability.
Title: Navigating the Limits of GenAI in Business Decisions—Beyond Productivity
Abstract:
Generative AI (GenAI) exhibits truly amazing capabilities, especially with data types characterized by high internal correlations, such as text, images, videos, and audio data. However, its efficacy encounters limitations when dealing with tabular data due to the lack of intra-dimensional correlations within such datasets. This poses a serious drawback in mission-critical business applications, as most important business decisions in today’s digital age rely on some sort of structured tabular data.
That’s why GenAI primarily functions as a productivity tool today. Despite the massive efficiency gain it offers, there aren’t many business applications of GenAI in mission-critical decisions that can directly drive profitability. While the profit equation is simple: Profit = Revenue – Cost, its maximization across industries is very complex, especially under the current volatile macroeconomic conditions. Successful AI profit optimization requires the harmony of 3 AI-related technologies. This session will also introduce these innovations and show how they can be seamlessly integrated with GenAI to drive the next level of profitable growth.
Guest Lecture at Sofia University – Faculty of Mathematics and Informatics
While BAS (established in 1869) serves as Bulgaria’s preeminent scientific institution, focusing on advancing scientific knowledge through fundamental and applied research, it doesn’t have an undergraduate student body. In contrast, Sofia University (SU, founded in 1888) is Bulgaria’s oldest higher education institution, providing comprehensive undergraduate and postgraduate education across diverse disciplines.
As I mentioned earlier, PROS has a sizable engineering team in Sofia, and quite a few of them are alumni of SU from the Faculty of Mathematics and Informatics (FMI). Through my esteemed colleagues’ connection to SU and our HR team in Bulgaria, I am very privileged to be giving a guest lecture at SU to the FMI.
For those who have known me for a long time, you know I always enjoyed speaking to undergraduate students, as they are not afraid to ask the obvious questions. Because sometimes, those seemingly innocuously naive questions often turned out to be much more complicated. For example, the famous Feynman’s Reverse Sprinkler Problem is only solved this January, after 140 years. If you are interested, see this video for an accessible exposition.
Since many of the students at SU-FMI are future software engineers, I think it would be a very relevant topic to discuss the impact of GenAI on software development. GenAI can be a liberator that frees coders from doing mundane coding tasks (e.g. writing test cases, commenting, and documentation, etc.). But could it one day turn into a job terminator that eliminates the need for engineers altogether? This is still a topic of heated debate. Although no one can predict the future, we can certainly entertain this possibility by examining how enterprises are using CodeGen tools today.
Title: The 2 Cutting Edges of GenAI: Coding Liberator or Job Terminator?
Abstract:
The capacity of Generative AI (GenAI) to generate code, debug programs, and write documentation poses both opportunities and threats for software developers. Will GenAI liberate coders from mundane tasks, or will it eliminate engineers from the job market? To answer this question, we need to delve into the inner workings of large language models (LLMs) like ChatGPT.
This session will show you how LLMs work, how they are trained, and why they hallucinate. Despite the hallucinatory nature of all GenAI, we will provide a practical technique to use ChatGPT safely and productively in the enterprise setting. Beyond ChatGPT, we will uncover the magic behind GenAI’s creative potential—from engaging dialogues to beautiful artwork and captivating tunes, and we will understand why GenAI works so well. Meanwhile, we will inspect the code-generation (CodeGen) use case, discuss its implications, and examine how enterprises use CodeGen tools today. This interactive learning experience will equip you with deep knowledge of GenAI to use it responsibly for software engineering and beyond.
Conclusion
BAS and SU are 2 of the oldest and most prestigious academic institutions in Bulgaria. They contribute significantly to Bulgaria’s academic and intellectual landscape, albeit with distinct roles and missions. I am deeply honored to be lecturing at both institutions, albeit with slightly different emphases. If you would like to attend, REGISTER HERE.
Beyond the lecture halls, I am also excited by the prospect of connecting with our engineering teams in Sofia. Moreover, I am eager to meet like-minded people and forge new relationships with the local academic and business communities. Because this journey is not merely about knowledge sharing. It’s a celebration of chance, curiosity, innovation, and the power of passionate collaboration that transcends borders and institutions.
Throughout my life, I’ve been fortunate to be frequently reminded to “never doubt that a small group of thoughtful, committed citizens can change the world.” The rest comes down to finding those passionate, like-minded, and committed individuals. Remember,… this journey started with the serendipitous encounter with one such individual – Prof. Iliev. Yet, I think this is just the beginning, with many more to come. I hope to see you in April, in person, in Sofia.