Interview

AITech Interview with Victor Thu, President of Datatron

Datatron

Victor Thu talks about his favorite part about working in the industry and how he manages to stay relevant in such a fast-paced industry.

1. Can you tell us more about yourself and your career before Datatron?

I have specialized in product marketing, go-to-market and product management in C-level and director positions. Prior to my current role in startups, I had multiple leadership roles, including a 3-year Asia Pacific + Japan posting, in product management and product marketing at VMware and Citrix.

Over the last six years, my focus has been on machine learning startups and how they can help large enterprises, including banking and financial services, international airports, retailers and manufacturers to solve crucial business challenges with advanced AI. A common challenge all the companies I’ve worked with, large and small, are still having is how to get AI models out of the lab and into production at scale, which has led me to Datatron.

2. Your major background is in product advertising and marketing. How did you get your start in the field of ML and AI?

I have always enjoyed translating complex tech topics into ideas that people can relate to.  My close friends even refer to me as the technology-whisperer.  So in many of my previous roles, I enjoyed learning about new technologies and getting to the heart of why such technologies matter to people.

Once I was at VMware watching a keynote presentation by a famous Standard AI professor, Dr. Fei-Fei Li. Dr Li’s keynote presentation was so captivating that it served as a turning point for me in my career. That presentation convinced me that I wanted to be part of the next wave of technology where we can use AI/ML to solve business challenges.

Since then, I have been with a number of AI/ML startups who were on the front lines of creating technology to address real enterprise pain points. I was fortunate enough to be at a number of startups where I worked very closely with Ph.D-level ML scientists. My time with these companies taught me a tremendous amount, especially in AI/ML. And since the space is ever-changing, I’m still learning everyday.

So it is truly a passion for technology and how to leverage it to help others that brought me to working closely in AI/ML.

3. Can you tell us more about Datatron? What are the technology solutions that Datatron offers?

Datatron helps enterprises operationalize their AI/ML models at scale. Designed for major retail, manufacturing, finance and pharmaceutical organizations, Datatron delivers fully automated operationalization of AI/ML models for testing, validation, deployment, and scale without the need for an army of expensive engineers or solution lock-in. 

Today, there’s no doubt that data scientists can simply push models into production with relative ease. However, when it comes to getting it to scale, implement A/B testing and validation, monitoring, and more, Datatron becomes a necessity, as those are not trivial.

Datatron is platform-agnostic and works well with any ML development platform on the market, including open-source tools, automating the entire process from model containerization to operationalization. Because Datatron delivers the operationalization capabilities for AI/ML, we give enterprises and data scientists tremendous flexibility to build models the way they were intended to be built, enabling models to be deployed into production easily and scaled rapidly. 

Because the Datatron solution manages the infrastructure supporting these models, DevOps teams no longer have to spend precious time trying to figure out how to support AI/ML models once deployed and are able to train models on real-world data more rapidly, speeding ROI and TTV. 

4. Datatron focuses on MLOps. Can you elaborate on what is MLOps and its features?

MLOps is essentially codifying and simplifying the highly artisanal process of getting AI/ML models from prototype to production.

One of the biggest misconceptions today is that once data scientists have built their AI models, they can get them out into production.
However, the reality is that it can take up to a year before a model can be deployed at scale.

The main reason is that people who have expertise in developing models do not have software engineering expertise. An analogy of this is architects who design skyscrapers are not developers who construct them.

So MLOps is essentially the bridge between model developers and software engineering. Instead of having to spend over 12 months to get models out to production at scale, MLOps can now cut that down to days.

MLOps really is the way to help AI/ML models into production while preserving resources and scaling down on cost. We are a sustainable source for successful AI/ML models and serve as a catalyst for revenue. Companies can finally get their ROI from their AI and ML projects. Regardless of your margins, you can produce results using MLOps.

5. What’s your favorite part about working in the industry?

My favorite part about the tech industry is seeing the constant innovation and how they could change the life of many.

I still remember a video clip from the Today’s Show in the late 90s where the anchors were questioning what the Internet was about.

This is where we are with AI/ML, perhaps even slightly earlier than the late 90s when the Internet was becoming a hot topic.

To see how technology is transforming the life of everyone for the better is truly amazing and I cannot stop geeking out about it.

6. Tell us a bit about your culture. What makes Datatron’s culture unique?

Datatron’s culture is that of customer-centric innovation. We have some of the top engineers in the industry and we focus on delivering innovation that matters to our clients. Our entire team is super focused on helping our clients and working to deliver solutions that can help make their lives easier.

A lot of companies talk about their open-door policy and integrity. This is not a marketing motto for us – we truly believe in these aspects of culture. My team often hears me telling them to share bad news or challenges with me so that we can address them honestly and correctly as a team.

7. Could you give a sneak peek into the recent developments at Datatron?

Datatron recently announced the latest version of our enterprise-grade MLOps platform. Updates include increased flexibility, a new interface that simplifies data scientists’ workflow, and ease-of-use enhancements for the operational teams, resulting in an additional productivity gain of up to 68%. 

Datatron Version 3.0 allows enterprises to achieve results with AI and ML by removing the roadblocks that prevent successful deployment. Based on feedback from clients, we are further simplifying the life of data scientists, making it easier to register, iterate and deploy models with just a few simple commands

Our recent partnership with Spectro Cloud is exciting and is helping us drive this simplification forward. Through the partnership customers gain a simpler way to deploy and manage Datatron’s Kubernetes infrastructure. In addition to supporting AWS, the new capability extends this greater simplicity to Google Cloud Platform and Microsoft Azure. This improved deployment and management capability removes the complexities of enterprises having to learn and manage Kubernetes.

8. In your opinion, what are the most exciting topics in AITech right now? How do you keep up with the constantly changing landscape?

It is indeed exciting in the AI world. My personal favorite is computer vision as there are tremendous possibilities with computer vision and yet we have barely even scratched the surface of it. There’s also a lot of misconceptions and fears about computer vision. The use cases are just beginning to develop and they are incredibly exciting.

The next element that I’m watching carefully is what is broadly called TinyML. Some of the AI models, including the latest NLP models and computer vision models, are so large today that you need to have significant network connectivity and expensive compute resources for inferencing. What if you can get the same inferencing capabilities on a small device with little or even no network connectivity? That’s an amazing thing to have.

9. What are Datatron’s plans for expansion and growth? Where do you see it in the coming years?

We have seen an acceleration of companies understanding the need to leverage an MLOps solution like Datatron in order to be successful. I’m even willing to venture to say Datatron is shepherding MLOps 2.0 whereby we help democratize resources that are typically only available for large enterprises or tech giants to allow small and medium companies to have the same cloud-scale tool to deploy their AI/ML models at scale with just a few engineers!

So we continue to focus on building up the team to enhance our platform that enables businesses to get their AI/ML deployed at scale at a fraction of the cost.

We know very clearly where our core capabilities are and we will continue to make sure we are the best in delivering that for our customers.

10. How do you stay motivated? What are your key learnings from your career so far?

Seeing how customers are successful with our product is a huge motivator for me. Especially solving their complex challenges and knowing that they have a lot of other options but chose Datatron to help in their journey.

Working with a very smart team of engineers is also very motivating for me. Engaging in discussions and debates; challenging each other on what can or cannot be done then coming up with a “light bulb” moment is extremely exciting.

One key thing I have learned from other tech leaders in the business is that companies are here to create products that benefit their customers. I am a big believer in that. If the product is offering value to your customers, the shareholder value will follow.

Another is how leaders in a company need to be grounded and willing to have open and honest discussions, facilitate communication and engage in debate with their teams.

11. What movie/book has inspired you recently?

A book that has inspired me recently is Extreme Ownership by Jocko Willink , Leif Babin. The book is about how we need to take ownership, or even extreme ownership, of what is happening in our life or at work. Even though some responsibilities are not under your control, as a leader, you are ultimately responsible and need to identify ways to work with the team to achieve the ultimate goals. It forces you to think about what it takes to reach each outcome without resorting to blaming others. This is a characteristic that is missing from a lot of leaders today.

12. What is the most significant piece of advice you would want to give to company leaders?

I think this is why the USA is such a unique place for businesses – we get to learn in real time. We see how large companies disappear over time because of the leaders becoming complacent and becoming risk averse.

Leaders must be willing to take giant calculated risks on a regular basis. Especially in the hyper connected world today where changes happen fast and furious, the notion of “if it ain’t broke, don’t fix it” will actually end up killing your company. The competition will do the fixing for you!

Visit AITechPark for cutting-edge Tech Trends around AI, ML, Cybersecurity, along with AITech News, and timely updates from industry professionals!

Victor Thu

President of Datatron

Victor Thu is president of Datatron. Throughout his career, Victor has specialized in product marketing, go-to-market and product management in C-level and director positions for companies such as Petuum, VMware and Citrix.

Related posts

Interview with Dr Radu Rusu, Co-founder and CEO at Fyusion

AI TechPark

AITech Interview with Michel Spruijt, CRO, Brain Corp

AI TechPark

Interview with Eyal Arazi, Security Product Marketing Manager at Radware

AI TechPark