AI might have crept up on us while it was on the rise. But Arvs tells us he already caught wind of some signals very early on.
“Rather than wait around for someone to say the same or even tell me what to do about it, I figured I should get to work,” he shares.
Tucked away in his home office, he tinkered with a managed machine learning service called Amazon SageMaker for months, performing a series of experiments and deployments with algorithms and techniques used for AI and ML projects.
The end result was what he refers to as a “cookbook” with relevant and proven recipes meant to help data scientists and machine learning engineers working on their next big thing.
As he handed in his manuscript to his publishers, they told him it was a few hundred pages too long, which then became the foundation of his second book in two years.
As if that wasn’t impressive enough, all of this happened during a global pandemic that was transforming everything we knew about the world.
While most of us were struggling to find our footing, Joshua Arvin Lat was running at breakneck speeds, successfully matching the pace of the AI wave.
Experimenting as part of everyday life
As the chief technology officer of NuWorks Interactive Labs, Arvs is expected to delegate the trial-and-error process to specialists on the team. After all, his existing workload is already heavy enough as it is:
“As CTO, around 30% of my time is allotted for collaboration: governing the processes and standards of the whole organization, which include management, security, hardware and software subscriptions, and budgeting. Then, maybe 70% goes to the actual tech and product development work: generative AI, engineering, automation, experience design, creative tech, mobile and app development. It really is a little bit of everything.”
But as a leader, Arvs is a firm believer that management skills need to be backed up with practical technical expertise.
“If you don’t experiment much earlier as a leader, you’ll have a hard time managing projects and initiatives because you don’t know what’s happening behind the scenes,” he shared.
This is especially true in the ever-expanding landscape of generative AI: “Anyone can understand the concepts given enough time, but hands-on skills, including the ability to troubleshoot and manage failure – that’s hard to find.”
Arvs’s high-functioning, high-achieving nature can be traced back to his time at the Philippine Science High School, where the system’s four-year focus on coding and technology hardwired his brain for the better. “I remember creating a strategy game similar to the Pokemon and Yu-Gi-Oh trading card games, where players duel against an AI version of me: that was my first brush with AI,” he recalls.
“As a student, I didn’t care if it was hard or if I was getting value for my money, I just wanted to make the best [game] I could.”
True enough, he was able to do just that after “around two weeks to a month”, creating an AI-powered game that could compete against, outmatch, and frustrate his peers.
After a temporary detour at the University of the Philippines’ College of Architecture, Arvs returned to where his heart truly was: computer science. During his thesis year, he stumbled upon an opportunity to submit his team's cybersecurity paper and defend it against other researchers. His three-man team bested representatives from across the world, including Europe and the States, many of whom were master's and Ph.D. holders, winning first place overall in the Kaspersky IT Security for the Next Generation global cup.
.webp)
Unlearning and reworking through writing
Since then, Arvs has landed stints in local and international groups of startups and cemented his status as an AWS Machine Learning Hero. Holding talks about engineering systems as well as balancing management skills with technical know-how carried a surprising fringe benefit for him:
“It ended up being one of the main reasons that inspired me to write the book: after speaking with a lot of professionals and diving deeper into their work challenges, I was able to better understand their needs and how I could fill in the gaps in their knowledge,” he says.
Book #3, released in October of last year, focused on cloud security and penetration testing – something he foresaw as a priority area after the rise of AI. While he was originally planning to have his fourth book on serverless security as his last one, he felt compelled to create a second edition of his second machine learning book to make way for the latest advancements in Generative AI. So yes, to address the elephant in the room: Arvs is writing two books. At the same time.
“The biggest challenge for me in writing is the rapid change of technology: I had mentioned in my AI in Action talk that as an author, you will have to update the contents of the books you wrote before because the rate of innovation in the field is moving so quickly. I was practically forced to write a second edition,” he says.
Of course, there’s also the risk of boring readers with the nitty gritty that’s just a quick Google search or YouTube tutorial away. “When writing, I found that it shouldn’t be all technical: there should be sections devoted to certain concepts, that segue into real-life examples and scenarios by companies before providing space for the reader to apply what they just learned,” Arvs shares.
This is why he crafts and tests all of his solutions himself, with the firm resolve and natural savvy of a textbook mad scientist.
Cultivating curiosity within teams
Arvs’ knack for experimentation has also proven necessary at work, where the percentage of AI projects has increased exponentially over the past few years. From working on simple, rule-based chatbots, Arvs’ team has built entire machine learning-powered systems meant to reduce the time for manual processes from two days to two seconds.
“When developing generative AI solutions and tools, developers should exercise caution when using internet resources and examples as they may be outdated, incorrect, and in some cases could introduce security vulnerabilities in the applications,” he says.
“We’re always operating under the mindset that clients have paid us to give them a working solution so we need to be aware of its limitations and requirements.”
Arvs also encourages his team members to be independent self-starters: even those who aren’t formally trained in research are expected to perform basic R&D tasks so they can all contribute to an initial proof of concept. Laying these foundational skills seems to be working in their favor: “Based on our track record, we’ve had very minimal projects that have encountered issues; we also get to keep our work culture healthy because we barely see resignations from team members,” he boasts.
As human beings with a finite attention span and capacity for work, it’s hard to believe that we all have the same 24 hours as Arvs. It’s almost 7pm on a Friday night when the call draws to a close, but it seems like his day is just beginning. “Now is actually the best time to write because I don’t have any work to think about the next day: I can just go at it until the early hours of the morning,” he beams.
When Swarm asks him where he gets the energy to maintain his routine, he reveals his secret:
Boxing and muay thai during the weekends,
with his girlfriend.
Talk about developing a strong body, mind, and relationship.
“It’s really those high-intensity exercises that help me gain energy and extend the prime years of my life. I’ve mastered the technique on how to recover easily,” Arvs shares.
All these things considered, it seems like Arvs has cracked the code to flourishing personally and professionally. But on the off chance that he gets something wrong, he of all people knows that there is nothing to lose in starting from scratch and trying again.
Resources
[Article] Kaspersky Lab Announces the Winners of Its Annual Student Conference (2012)
