Transformation of the Week
Vincent Kienzler

Transcript

Lusk: Hey, Vincent, thanks for joining me today. Congratulations on winning the Transformation of The Week. Tell me a little bit more about your background.

Vincent: Yes thanks, Chris, pleasure to be here. And thanks for, inviting me over. So I started as a computer science engineer a while back, and I worked initially as a telecommunications engineer, working on satellite projects. But somehow I felt that I wanted to learn more. I was curious, so I went back to school, studied a bit of economics and, philosophy, and, then started a couple of companies, one is still up and running.

The other one, they successfully exited. Somehow along the way, I realized that tech was really my first love. And I went back into tech, ore less eight years. I did some hardware development, software application development and data, and more recently for the last three years or so, went into AI. I am currently working as a lead tech venture builder for an investment fund. It’s basically I am a CTO in that fund and I run technology due diligence on the companies we invest in. And once we have invested in those companies, I lead technical projects to help them support those companies to address the gaps that we identify in due diligence. And, sometimes they cover some roles of fractional CTO, CIO, CPOAs and the like.

Lusk: So what got you interested in Gen AI in the first place?

Vincent: Two things, actually, the first one is really curiosity and the “wow” effect. I started playing with large language model some years back, and I realized this is really powerful, and I’m pretty sure that this is going to be the next, big tech thing. And I really wanted to understand what was happening behind the scenes.

So that’s one of the reason why I went into it. The other reason is a bit more strategic. I realized that it could lead to a lot of opportunity. I have had in the back of my mind starting a third company. And I really thought that AI could be one thing, one option, and even if it’s not an AI company, AI would be part of it.

Lusk: Now, you graduated from our very first AI Engineering Bootcamp, you graduated about ten months ago. Have you been able to put anything into practice, whether it’s at work or on any personal projects?

Vincent: Yes, absolutely. I’ve been basically using AI every day, since then. And, at work, on my regular work, I’ve been implementing two different projects. One is, cybersecurity advisory boards that uses RAG to collect example of cybersecurity incidents, you know, to answer questions. So for the background, as part of my work, we are helping companies become more secure. So we are doing cyber security audits of their practices, and we are also conducting trainings. So as an add-on to those trainings, we offer these, cybersecurity chatbots to help them answer any question that they might still have. So this is one thing that I implemented following the bootcamp.

The second thing is I started working as a consultant for an AI startup that’s working on setting up the go-to-market web application that helps salespeople close deals. So I’m working on all the AI engine to get that up and running.

As part of our technical due diligence work, we conduct interviews, we collect information and write investment memos. So I’ve been using AI to automate parts of that process. I am now submitting Google Forms to the companies we invest in and collecting that information through Google Forms, and then using AI to write parts of the investment memo, the parts where you get a summary of the state of those companies. I’m definitely using it, and I’m using it on a regular basis.

Lusk: Now, you’ve been working on Gen AI solutions for quite some time now. What’s one big challenge that you’re currently trying to solve?

Vincent: Actually, I have two big challenges I’m working on. The first one is privacy and the taxes. You know, when you work on enterprise AI, you are working with data that comes from different sources, Slack, emails, Google Docs or any type of docs, and so on. And you try to put all that data together to provide recommendation or answers to the user questions. The problem is that sometimes you might inadvertently get some information that’s confidential into those answers. And that’s a big problem, and that’s a big no for large companies. One of the things have been working on is how can we make sure that the user that is using AI has access via that AI only to the information that you should have access to and not to confidential information that, you shouldn’t see.

The second piece is around issuing recommendation based on an analysis of gaps in investment analysis. It’s quite good to use AI for generate summary and to synthesize data that you have about the company. But the issue is when you ask your AI to generate a recommendation or identify gaps, that’s where answers are not automatically of a good quality. So this requires a bit more work. This requires fine tuning, playing with the RAG sometimes, and finding the right combination of RAG and fine-tuning to get good quality answers.

Lusk: Now you’re out there building, shipping, and sharing every single day, and it seems like nothing gets in your way. What words of wisdom do you have for people who are considering Gen AI upskilling?

Vincent: One is be prepared to keep learning as AI is evolving very quickly, the library that you are using now will be likely very different in four months. Voice is becoming a thing. Videos are becoming a thing as well. So you really need to keep learning.

That was the first one, the second one is be hands on test and be ready to fail. It’s by failing that you are going to eventually be successful. You need to get used to how AI answers your prompts. You need to get used to the differences between RAG and fine tuning, when to use what for what purposes. Different language models have different personalities. By using them, testing, failing, you will get a sense of what’s their personality. What’s their tone, what they are good at. And you will be able to use the best one for the best use case.

Lusk: Vincent, congrats again on winning the Transformation of The Week, and thanks for joining me today. Where can people connect with you?

Vincent: So the best way to connect with me is on LinkedIn -Vincent dot Kinsler – and follow me on GitHub as well.