The Tech Trek

Allan Leinwand, CTO at Webflow, joins me to explore how AI is reshaping engineering workflows, from code generation to team structure. We dig into how AI tools are boosting productivity, enabling faster onboarding for junior engineers, and freeing up senior talent to focus on distributed systems and business-critical challenges. Allan shares real examples of automation in action, how his team measures success, and why the future of software engineering will be even more dynamic than its past.

Key Takeaways

• AI-powered tools like code generation and multimodal debugging are changing how engineers interact with code

• Junior engineers can now ramp up and make meaningful contributions faster than ever before

• Senior engineers are moving closer to the business by tackling architectural and scalability problems

• Automation is cutting down repetitive tasks, increasing flow time, and boosting ship rates

• AI is influencing not just engineering, but also product workflows and even how methodologies like Scrum might evolve

Timestamped Highlights

01:45 – Inside Webflow’s AI-powered engineering stack and tools every developer gets

03:57 – How AI is shifting the engineer–code relationship from typing to prompting and reviewing

07:15 – Why junior engineers are thriving in the age of AI

13:37 – Senior engineers focusing on distributed systems and architectural challenges

16:15 – Automating “paper cut” bug fixes with AI agents and background processes

21:09 – AI’s role in expanding software creation to non-engineers and influencing product workflows

Quote of the Episode

“The relationship with code is changing. We can talk to the code base, use AI to fix bugs, and still have humans in the loop to make sure it’s the right answer.” — Allan Leinwand

Resources Mentioned

• Webflow — https://webflow.com

• Webflow Forums — https://forum.webflow.com

Call to Action

If you found this conversation valuable, share it with another tech leader who’s navigating AI adoption. Follow the show for more insights from engineering leaders shaping the future of work.

Nikola Borisov, CEO and co-founder of Deep Infra, joins the show to unpack the rapid evolution of AI inference, the hardware race powering it, and how startups can actually keep up without burning out. From open source breakthroughs to the business realities of model selection, Nikola shares why speed, efficiency, and strategic focus matter more than ever. If you’re building in AI, this conversation will help you see the road ahead more clearly.

Key Takeaways

• Open source AI models are advancing at a pace that forces founders to choose focus over chasing every release.

• First mover advantage in AI is real but plays out differently than in consumer tech because models are often black boxes to end users.

• Infrastructure and hardware strategy can make or break AI product delivery, especially for startups.

• Efficient inference may become more important than efficient training as AI usage scales.

• Optimizing for specific customer needs can create significant performance and cost advantages.

Timestamped Highlights

[02:12] How far AI has come — and why we’re still under 10% of its future potential

[04:11] The challenge of keeping pace with constant model releases

[08:12] Why differentiation between models still matters for builders

[14:08] The hidden costs and strategies of AI hardware infrastructure

[18:05] Why inference efficiency could eclipse training efficiency

[21:46] Lessons from missed opportunities and unexpected shifts in model innovation

Quote of the Episode

“Being more efficient at inference is going to be way more important than being very efficient at training.” — Nikola Borisov

Resources Mentioned

DeepInfra — https://deepinfra.com

Nikola Borisov on LinkedIn — https://www.linkedin.com/in/nikolab

Call to Action

If you enjoyed this conversation, share it with someone building in AI and subscribe so you never miss an episode. Your next big idea might just come from the next one.

Zach Lloyd, CEO and founder of Warp, joins The Tech Trek to unpack what it really takes to build tools that transform the developer experience. From rethinking the terminal to balancing product focus with user growth, Zach shares hard-earned lessons from scaling products that developers actually want to use. This is a conversation about building with empathy, understanding workflows, and making deliberate trade-offs that move the needle.

Key Takeaways

• Why deep focus on the developer workflow leads to products that stick

• The importance of balancing big-picture vision with small, iterative improvements

• How to make trade-offs between growth experiments and core product quality

• Why some of the most powerful product ideas come from rethinking “old” tools

• The role of design and speed in shaping developer adoption

Timestamped Highlights

[03:15] The inspiration behind Warp and why the terminal needed rethinking

[09:42] Balancing user requests with long-term product vision

[14:10] How small quality-of-life improvements can have outsized impact

[21:55] Deciding when to invest in growth versus core product work

[28:30] Lessons from building for an audience of highly opinionated users

[36:05] Why the future of dev tools will blend speed, design, and collaboration

Quote of the Episode

“The best products come from understanding the real workflow pain and then removing it in a way that feels almost invisible to the user.”

Resources Mentioned

Warp: https://www.warp.dev

If you enjoyed this conversation, follow The Tech Trek on your favorite podcast platform and connect with me on LinkedIn for more insights from the leaders shaping the future of technology.

Sek Chai, CTO and cofounder of Latent AI, joins The Tech Trek to talk about what it actually takes to get AI running on the edge. We explore the real-world constraints of power, compute, and hardware diversity, why an agent-assisted workflow can accelerate MLOps, and how to choose models that are good enough to ship. Sek also breaks down lessons from selling into the federal market and explains why a clear guiding principle beats chasing every shiny opportunity.

Key Takeaways

Edge AI is a different game than the cloud. Power limits, hardware diversity, and deployment realities have to shape the design from day one.

The best model is the smallest one that delivers the capability and latency you need. Bigger isn’t always better.

An AI agent that understands your data, model, and hardware personas can move teams from idea to deployment much faster.

Whether you’re selling to federal or commercial buyers, lead with capability, then meet security and compliance needs.

A strong tenet should guide product direction and market focus more than raw market size.

Timestamped Highlights

00:30 Why edge optimization matters and what Latent AI does

01:09 The messy reality of heterogeneity and power constraints in edge deployments

02:54 Why most edge AI projects never ship and how an agent can change that

05:03 Mapping MLOps personas and tailoring the workflow for each

11:49 Selling to both federal and commercial buyers without losing focus

15:55 Building a company around a tenet rather than chasing every market

Quote of the Episode

“It’s not the model that you’re really chasing after. It’s that capability.”

Pro Tips

Define capability and constraints first—latency, frame rate, and power budget—then pick and optimize the model.

Collect and use telemetry from experiments and deployments to guide model and hardware choices.

If federal markets are in play, bake security and compliance into your early prototypes.

Call to Action

Enjoyed this episode? Follow The Tech Trek, rate us on Apple or Spotify, and share it with someone working on an edge AI project.

Berit Hoffmann, CEO and co-founder of Korl, joins The Tech Trek to share her candid journey from big tech leader to late-stage startup founder. With a resume that includes Google, Dell, and Sisu, Berit could have landed any top role—but she chose the riskier path of building her own AI company while raising two kids and fundraising while seven months pregnant. In this episode, she opens up about the internal tug-of-war, the realities of balancing family and founder life, and how she’s navigating the fast-moving, hype-driven world of AI. If you're a tech professional wondering when—or whether—to make your own leap, this one’s for you.

Key Takeaways:

Experience doesn’t remove fear—but it can sharpen your confidence in taking big risks

AI founders must constantly recalibrate as models evolve and moats evaporate

The best startups fall in love with the problem, not the initial solution

You don’t have to wait for perfect timing—it might never come

Execution and clarity win over buzzwords in a crowded AI market

Timestamped Highlights:

00:44 — What Korl actually does and why it's different from other AI presentation tools

02:30 — Why Berit waited to found a startup and how early roles shaped her confidence

07:03 — The hidden opportunity costs and fears of starting later in life

11:38 — Her zero-to-one playbook: validate the problem deeply before writing a line of code

15:50 — Fundraising in the age of AI hype and navigating the balance between clarity and buzz

20:33 — How she processes new AI releases and adapts strategy without spinning out

24:45 — What it was really like to raise VC funding while visibly pregnant

30:11 — Her honest take on founder-parent balance: sometimes 80% has to be enough

Quote of the Episode:

“There’s still such a gap between what many AI tools promise and what they actually deliver. Closing that gap is all about execution—and that’s where startups win.”

Resources Mentioned:

Koral: https://www.getkoral.com

Connect with Berit on LinkedIn: https://www.linkedin.com/in/berithoffmann/

Call to Action:

Enjoyed the conversation? Follow The Tech Trek for more real stories from tech builders and startup leaders. Share this episode with someone who's debating their next leap—you never know what might spark them to go for it.

What does it take to lead analytics at a truly global scale? In this episode, Amir sits down with Anant Veeravalli, Global Chief Data and Analytics Officer at Media Brands (part of IPG), to unpack how he built and scaled a Center of Excellence (COE) that spans regions, brands, and disciplines. Anant shares the real-world challenges of aligning thousands of data professionals under one strategic vision—and why analytics is far more than just reporting.

If you're leading data, analytics, or transformation work inside a large enterprise, this one is packed with battle-tested insight on structure, talent, AI adoption, and the real work of enabling data to drive business value.

Key Takeaways

A COE must be built around client outcomes, organizational excellence, and scalable innovation—not just reporting structure

True analytics value comes from harmonizing data, tools, and talent while making insights accessible and actionable

A skills-first approach helps align talent to opportunity, enabling flexibility and specialization at scale

AI isn’t just a buzzword—it’s already reshaping content, audience segmentation, modeling, and competitive intelligence

Communication during change is everything. Transparency, context, and repetition are essential to alignment and trust

Timestamped Highlights

03:00 — Why Media Brands needed a global Center of Excellence for analytics

05:30 — How they approached organizational change and stakeholder education

08:50 — Mapping the COE into four key capabilities: growth, audience analytics, data science, and data engineering

11:50 — Delivering flexible analytics support across diverse clients and geographies

15:45 — How AI is driving faster insights, better segmentation, and creative automation

21:30 — The #1 thing Anant would do differently if starting over: communicate the why more consistently and directly

Quote of the Episode

“We overly underestimate the value of transparent communication. If people don’t understand why the change matters to them, they’ll never be aligned.”

Pro Tips

Invest early in an internal asset library to avoid duplicated effort and unlock speed

When hiring, prioritize specialization over generalization—then connect specialists across a shared framework

Don’t just train on AI tools. Raise the entire organization’s AI literacy

Call to Action

Enjoyed this episode? Follow The Tech Trek for more conversations with leaders building the future of tech, data, and innovation. Share this episode with someone navigating data transformation—or connect with Anant Veeravalli on LinkedIn to keep the conversation going.

What if your phone didn’t need to hold your data at all? In this episode of The Tech Trek, Amir sits down with Jared Shepard, CEO of Hypori, to explore how virtualization at the edge is transforming security, mobility, and data ownership. Jared breaks down Hypori’s secure virtual mobile OS, originally built for the Department of Defense, and how it’s now entering the enterprise and consumer spaces. From eliminating mobile device management to protecting sensitive data from AI exposure, this conversation is a wake-up call for any tech leader thinking about security at the edge.

Key Takeaways:

Hypori’s virtual mobile OS allows users to access enterprise data securely without storing it on their device.

Virtualization collapses the attack surface by removing the edge device as a security risk.

U.S. enterprises prioritize convenience and security, while Europe pushes privacy due to GDPR—Hypori bridges both.

AI will soon enhance Hypori's platform through predictive resource allocation and network optimization.

The military’s extreme security standards helped Hypori harden its platform far beyond typical commercial use cases.

Timestamped Highlights:

01:30 — What Hypori is and how it turns any device into a secure, data-less terminal

05:30 — Real-world BYOD use cases, from consultants to GDPR-compliant European enterprises

11:20 — How virtualization changes the AI risk equation and protects enterprise data from agentic threats

15:50 — Why cybersecurity should stop blaming users and start simplifying their responsibilities

18:45 — How virtualization shrinks the attack surface and simplifies network defense

22:59 — What it’s like building for the Department of Defense and how that shaped Hypori’s product

Quote of the Episode:

“Maybe it doesn’t have to be a company’s fight versus your fight for whose data belongs on your phone. What if we could just take that problem away?”

Resources Mentioned:

Hypori: www.hypori.com

Call to Action:

If this episode got you rethinking your mobile security strategy, share it with your team or your CIO. Subscribe to The Tech Trek for more conversations at the intersection of leadership, innovation, and real-world security.

Feross Aboukhadijeh, founder and CEO of Socket, joins The Tech Trek to pull back the curtain on software supply chain security, why legacy tools are failing, and what it really takes to build trust into modern development. Feross explains how Socket is tackling vulnerabilities most vendors can't even detect and shares why they made a rare early-stage acquisition—and how it’s reshaping their roadmap.

Whether you’re an engineering leader, security pro, or founder eyeing M&A moves, this episode offers sharp insights into product strategy, AI implications, and the real work behind the scenes.

Key Takeaways:

Socket proactively secures the software supply chain by detecting malicious code injections and not just known vulnerabilities

Legacy tools rely on outdated databases and can’t keep up with real-time threats or malicious actors

The explosion of AI-generated code is expanding the attack surface and introducing new vectors like “slop squatting”

Socket’s acquisition of Kawana was driven by tight product fit, culture alignment, and shared technical DNA—not just business rationale

Reachability analysis reduced Socket’s security alert noise by 80 percent, boosting signal and developer trust

Timestamped Highlights:

01:00 — What Socket actually does and why open source dependency risk is a blind spot for most companies

06:40 — Why most tools in this space haven’t solved the real security problem—and how Socket is different

11:50 — AI’s unexpected impact on software security and the rise of hallucinated packages

16:30 — Behind Socket’s acquisition of Kawana and how academic research drove product synergy

22:58 — How integrating the acquisition is evolving Socket’s roadmap and deepening its technical edge

25:00 — What Feross learned from the legal side of M&A and how his past experience at Yahoo helped shape this one

Quote of the Episode:

“We care way more about first-party code than third-party code, even though it all runs in one app. That has to change.”

Resources Mentioned:

Socket: https://socket.dev

Call to Action:

Enjoyed the episode? Follow The Tech Trek to catch conversations with the builders shaping the future. And if you’re deep in security or scaling a dev team, check out socket.dev or reach out to Feross directly—he’s happy to share lessons learned.

What does it take to deliver innovation at just the right moment? In this episode of The Tech Trek, Amir sits down with Eric Hoffert, CTO at Kargo and former video leader at Apple and Spotify, to unpack the art and science of innovation timing. From building QuickTime at Apple to launching video at Spotify a decade before the market caught up, Eric shares stories that blend conviction, timing, and deep tech insight. This episode is a must-listen for anyone thinking about where AI, video, and advertising are headed—and how to lead through the chaos.

Key Takeaways:

Innovation is a blend of vision, timing, and execution—being first doesn’t matter if the world isn’t ready.

AI is shifting us from an attention economy to an intention economy, transforming how video content and advertising are personalized.

The best tech products often emerge from the intersection of diverse disciplines, creative conviction, and platform thinking.

Timing mistakes are common—even industry giants miss the mark by years—but conviction keeps the momentum alive.

Future video experiences will be radically personal, possibly generated in real time based on your preferences.

Timestamped Highlights:

00:58 — What Kargo does and why art + technology is their core advantage

02:04 — The behind-the-scenes story of inventing QuickTime at Apple

12:50 — Why Spotify’s video ambitions in 2011 were 15 years ahead of their time

17:33 — Can advertising become seamless and actually helpful? The AI-powered opportunity

22:23 — Scene-level targeting and privacy-preserving personalization in video

26:49 — Eric’s 3 keys to innovating at the right time: see around corners, surf the wave, move fast

Quote of the Episode:

“We’re shifting from an attention economy to an intention economy—where you’re in the driver’s seat of what you watch, and how it's monetized.”

Call to Action:

If this conversation got you thinking about where tech is headed, share it with a fellow builder or product leader. Follow the show for more deep dives into the minds shaping tomorrow’s tech—and drop a comment to let us know what resonated most.

What if your infrastructure could predict demand before it happens? In this episode, Nilo Rahmani, CEO and co-founder of Thoras AI, breaks down how predictive scaling is transforming the Kubernetes landscape. With over a decade of experience in site reliability engineering, Nilo shares why the observability market is slower to adopt AI—and why that might finally be changing. If you're navigating the pressures of DevOps or building AI tools for technical teams, this conversation is a must-listen.

Key Takeaways

AI adoption in reliability engineering isn’t about replacing humans—it’s about reducing fire drills and enabling better decision-making.

Predictive scaling using ML can dramatically cut cloud costs and reduce latency—without compromising reliability.

DevOps teams remain cautious with AI due to the high stakes of downtime and the need for human-in-the-loop decision-making.

The best tools won’t just optimize infrastructure—they’ll increase engineer confidence and operational readiness.

Nilo's founder journey started with a thesis and became unstoppable once she “couldn’t unsee the better way.”

Timestamped Highlights

[01:02] What Thoras AI actually does—and how it tackles the double challenge of utilization and cost

[03:12] Why reliability engineering is a high-stakes, thankless job and how AI can change that

[08:54] Can AI fully handle outages at 2 a.m.? Why human-in-the-loop still matters

[13:22] The low-hanging fruit: where ML delivers value fast in infrastructure planning

[17:56] Increasing confidence, not replacing engineers—rethinking developer experience with AI

[24:38] Nilo’s founder story: from SRE to CEO, driven by a problem too obvious to ignore

Quote of the Episode

“I couldn’t unsee that there’s a better way. Using machine learning to make decisions in reliability engineering is the obvious next step.”

Resources Mentioned

Thoras AI: thoras.ai

Connect with Nilo on LinkedIn: linkedin.com/in/nilo-devops

Call to Action

Enjoyed this episode? Share it with someone who lives on-call or is building for DevOps teams. Subscribe on your favorite platform and leave a review—it helps more tech professionals discover the show.