The Tech Trek

Soham Mazumdar, CEO and co founder of Wisdom AI, joins the show to talk about what it really takes to build again after major exits at Facebook and Rubrik. We get into the mindset shift required for a new startup, the danger of relying too much on past playbooks, and how to stay grounded when expectations rise.

If you want a real look at repeat founder decision making, this is the conversation to listen to.

Key Takeaways

• The biggest advantage of being a repeat founder is the ability to attract talent and early believers, but it does not replace the need for fresh thinking.

• Pattern matching can help with people decisions but can block you everywhere else if you assume the past will repeat.

• Feedback can steer you or mislead you. The real work is separating patterns from outliers and understanding the motivation behind what someone says.

• Every new company pulls you back to zero. Past success does not win customers or validate your idea.

• Early career operators who want to build should leap sooner than later. Even a failed startup can shape a long career.

Timestamped Highlights

02:01 How building tactile and Rubrik shaped his approach to Wisdom AI

04:07 What actually drives someone to found a company after a giant exit

05:52 Why repeat founders must fight the urge to reuse old playbooks

10:06 How to course correct when your first instincts are wrong

13:38 The danger of reacting too fast or too slow to customer feedback

18:02 How expectations shift once you have a track record

24:42 Why Wisdom AI connected with his earliest experiences at Google

25:28 The advice he wishes someone had given him before startup number one

A standout line

“The world pulls you down to the ground fast. Whatever you think you are, a new company reminds you that none of it matters unless you execute.”

Practical advice from the conversation

• Do not treat feedback as instructions. Treat it as signal to study. Look for repeated patterns, not one loud voice.

• Approach every new company with a clean mind. If your old patterns do not match the new environment, abandon them quickly.

• Think of your career as a long arc. Early risks create unexpected doors later.

Closing note

If the episode gave you something to think about, follow the show and share it with someone who wants a real look at the founder journey. You can also join the community on LinkedIn for more insights and upcoming episodes.

In this episode of The Tech Trek, Amir sits down with Sadi Khan, Co-Founder and CEO of Aven, to unpack how technology can make capital fairer for everyone. Sadi explains how Aven is tackling one of the world’s biggest inefficiencies—the trillion-dollar burden of consumer credit card debt—and why the solution lies in reducing the cost of capital through innovation. This is a deep dive into building products that require not just engineering skill, but endurance, conviction, and a long-term mindset.

Key Takeaways

• Aven’s mission is to cut credit card interest payments in half by rethinking how consumers access and use home equity.

• True innovation often comes from solving inefficiency, not chasing market trends.

• Complex problems create strong moats when founders are willing to grind through technical and regulatory barriers.

• Founders should pick problems worth spending a decade on—pivot less, persist more.

• Product success depends on identifying your “axis” and going all-in on being the best at that one thing.

Timestamped Highlights

00:40 — How Aven’s hybrid credit card + HELOC model is lowering the cost of borrowing for homeowners

04:10 — The moment Sadi realized the cost of capital was a massive, overlooked problem

12:34 — Why most lenders haven’t solved this yet and how Aven’s approach differs

19:33 — Building what others couldn’t: how persistence and engineering precision led to breakthroughs

23:36 — Choosing execution risk over market risk and what it takes to stay with a problem long enough to solve it

37:47 — Why picking the right “axis” is how great companies build an unshakable moat

Memorable Line

“The only problems worth working on are the ones worth working on for a very long time.”

Call to Action

If you enjoyed this episode, follow The Tech Trek for more conversations at the intersection of people, impact, and technology. Subscribe on your favorite platform and share it with someone building bold ideas.

Svetlana Zavelskaya, Head of Software Engineering for Data Platform and Infrastructure at Quanata, joins the show to unpack what it really takes to make the “impossible” possible in tech. From re-architecting a startup codebase to scaling innovation inside an insurance giant, she shares how her team turns complex R&D challenges into production-ready systems. This conversation dives deep into engineering discipline, AI tool adoption, and why the next wave of insurance innovation is powered by data and software.

Key Takeaways

• Real innovation often means balancing speed with long-term architecture decisions

• AI coding tools are valuable for exploration but need governance and clear security guardrails

• POCs fail when expectations aren’t aligned, not because the tech doesn’t work

• Insurance tech is evolving fast through telematics and context-based data models

• Well-structured, well-documented code is still the foundation for scalable innovation

Timestamped Highlights

00:33 How telematics is changing the economics of insurance and rewarding better drivers

03:59 Cars as software platforms and what that means for data privacy and innovation

06:02 The growing pains of re-architecting an organically built startup codebase

08:38 Evaluating new AI tools and maintaining data security across teams

11:08 Why most AI POCs never make it to production

16:29 How Quanata’s R&D work feeds into State Farm’s larger technology initiatives

20:40 Safe-driving challenges, behavioral change, and saving lives with data

A Thought That Stuck

“If we can prevent just 1 percent of drivers in the world from using their phone behind the wheel, imagine how many lives we can save.”

Pro Tips

• Before starting a POC, define if it’s an experiment or a potential product foundation

• Let engineers explore new tools but build frameworks to govern how data and results are handled

Call to Action

If you enjoy exploring how data, AI, and engineering innovation come together to solve real-world problems, follow The Tech Trek on Apple Podcasts or Spotify and share this episode with a colleague who builds at the edge of what’s possible.

Nikhil Gupta, founder and CTO of Vapi, joins Amir to talk about how voice AI is reshaping the way we connect with businesses. From customer support to healthcare, Nikhil explains how voice agents can bring back the human side of digital interactions. This is a look at where real conversation meets real technology and what happens when machines start to understand us like people do.

Key Takeaways

• Voice AI creates genuine, human-like engagement instead of the usual scripted support.

• The next wave of AI will personalize relationships at scale while protecting privacy.

• Full duplex voice models will make conversations flow naturally and feel real.

• Businesses will use voice agents to understand customers, not just respond to them.

• Our phones and screens may evolve as voice becomes the primary interface.

Timestamped Highlights

01:08 — What Vapi does and how it reached 400,000 developers

02:15 — Why voice AI is one of the few areas showing clear ROI

06:09 — How AI can make customer relationships human again

11:18 — Building trust and privacy into voice-based systems

16:48 — Blending text, voice, and context into a single experience

19:05 — Rethinking our devices as voice replaces the screen

A moment that stands out

“Every person should feel like they can just text their hospital, and it knows exactly who they are, what they need, and when to help.” — Nikhil Gupta

Pro Tip

Start small. Use voice AI where conversation improves experience or clarity. It’s not about automation; it’s about creating connection.

Call to Action

Share this episode with someone exploring AI in their business and follow The Tech Trek for more stories about people, impact, and technology.

Alex Daniels, Founder and CTO at Predoc, joins the show to share how he is building a mission driven healthtech company that is changing how medical data is accessed and used. He opens up about the personal story that inspired Predoc, how he keeps culture authentic while scaling, and what zero turnover really looks like in a startup. From hiring philosophies to equity design to managing context switching, Alex brings a deeply human view of leadership in engineering.

Key Takeaways

• Building culture starts with personal connection. Founders who share their why help every new hire connect to mission and meaning.

• The best hiring filters are values and networks, not just tech stack alignment.

• Predoc’s culture formula of high agency, urgency, meritocracy, and transparency keeps turnover at zero.

• Equity is not just compensation. It is shared ownership and long term motivation.

• Flat structures and super ICs can scale effectively when leaders stay close to the work.

Timestamped Highlights

[01:30] How a personal loss and a lifelong heart condition inspired Predoc’s mission to fix healthcare data

[05:20] Inside Predoc’s culture formula and why it has helped them retain every hire for three years

[09:40] Why core values stay constant but merit evolves as the company grows

[13:00] Rethinking equity and risk for early startup employees

[15:10] How Predoc combats AI assisted interview cheating and keeps hiring authentic

[23:45] Building a flat team structure where directors are still super ICs

[30:00] Alex’s approach to managing context switching and mental decompression

Memorable Line

“We cared about what he cared about and why would he care about what we care about if I don’t care about him?”

Call to Action

If you enjoyed this conversation, follow The Tech Trek for more candid talks with founders and tech leaders shaping the future of engineering and culture. Subscribe on Spotify or Apple Podcasts and join the discussion on LinkedIn.

Jordan Tigani, CEO and cofounder of MotherDuck, knows what world class infrastructure looks like. He spent years building Google BigQuery before taking those lessons into the startup world. In this episode, he breaks down why building infrastructure products is fundamentally different from typical SaaS and why founders who don’t understand that difference are in for a painful surprise.

What You'll Learn

There are no shortcuts in infrastructure. You can’t just wire together existing open source components and call it a product. Real infrastructure requires contributing meaningfully to the state of the art, and that takes time, money, and deeper technical investment than most founders expect.

Starting with startups, not enterprises, is often the smarter play. Early stage infrastructure companies should target other startups first because they’re more comfortable with bleeding edge tech, have lower security barriers, and won’t force you to spend three engineers building custom auth instead of your actual product.

Scaling down is the new scaling up. Jordan saw pressure at SingleStore to make databases smaller and more efficient, not just bigger. That insight led to MotherDuck, which is built on DuckDB—a database that can run in a car, scale to massive cloud instances, and challenge the coordination overhead of legacy distributed systems.

Bottoms up engineering cultures win in infrastructure. At BigQuery, engineers close to customer problems could ship fast and independently. Jordan’s recreating that at MotherDuck by removing layers between engineers and customers, because creative problem solving requires understanding business constraints, not just technical ones.

Convincing people you can scale is half the battle. The best proof is customers who look like your next target and can vouch for you. Next best is real data and benchmarks. If you don’t have those yet, lean on implementation support and help prospects test at scale themselves. Early on, sometimes all you have is your word.

Timestamped Highlights

[01:22] Why infrastructure takes longer to build than typical SaaS products and why there’s no shallow way to do it

[06:57] The MVP dilemma: finding product market fit when enterprises demand reliability from day one

[11:44] Lessons from BigQuery and SingleStore—what to carry over from big tech and what to leave behind

[21:21] The gap in the market that led to MotherDuck: why distributed databases don’t scale down and why that matters now

[26:10] Redefining scale: why 100 users on one giant instance isn’t necessarily better than 100 auto scaling individual instances

[29:08] The hierarchy of proof: from customer testimonials to benchmarks to trust me, it’ll work

A Line to Remember

“If you really want to build an infrastructure product, you can’t just string existing components together. You actually have to contribute meaningfully to improving the state of the art.”

Stay Connected

If this breakdown of infrastructure startups resonated with you, subscribe so you don’t miss future episodes. And if you’re building in this space or thinking about it, connect with Jordan on LinkedIn. He’s committed to paying forward the help he got as a founder.

Ogi Kavazovic, co-founder and CEO of House Rx, joins the show to unpack what most product leaders miss about building for enterprise software. Drawing from two decades in tech, Ogi breaks down how product management shifts when you move from B2C or “B to small B” to true enterprise—what he calls “B to Big B.” He explains why traditional user research frameworks don’t hold up, how buyer research should actually be done through sales and marketing motions, and how to keep engineering teams aligned when the product takes years to build.

Key Takeaways

• Building for enterprise (B to Big B) requires selling to buyers and users—two very different audiences with distinct needs.

• Buyer research is not user research—it happens through early sales decks, vision slides, and iterative storytelling that test how well a concept resonates before code is written.

• Pre-selling a “fantasy product” through slides helps validate the market fit and shapes the first version of your product strategy.

• Engineering for enterprise software demands simulated iteration—testing features internally long before the MVP is complete.

• Vision alignment between product, marketing, and engineering is crucial to avoid two-year build tunnels and ensure team motivation.

Timestamped Highlights

[03:12] The overlooked divide between B2B and true enterprise—why “B to Big B” changes everything for product teams.

[10:47] How buyer research actually works and why it starts with slides, not software.

[17:40] The difference between pitching VCs and pitching enterprise buyers—and why they care about totally different things.

[22:29] The engineering challenge of building massive enterprise systems and why agile methods fall short.

[30:11] How to keep teams motivated and moving forward when the product roadmap spans years.

Standout Moment

“You can pre-sell a product before it even exists. That sales and marketing artifact—the deck you built to sell your vision—can become the blueprint for your product strategy.”

Pro Tips

Start with conversations, not code. Use early customer and buyer meetings to validate your story through slides, then hand your engineers a vision they know can sell.

Call to Action

If you enjoyed this episode, share it with a fellow product leader or founder navigating enterprise challenges. Follow The Tech Trek for more conversations that connect people, impact, and technology.

Astasia Myers, General Partner at Felicis, breaks down how venture capital is betting on AI and why over 80% of their recent investments are in this space. But this isn’t just another “AI is the future” conversation. We dig into the real ROI happening right now in healthcare voice agents, why MIT says 95% of GenAI projects fail to reach production, and what needs to happen for that number to flip. If you’re building, investing, or just trying to understand where enterprise AI is actually working (not just hyped), this episode cuts through the noise.

What You'll Learn

The labor replacement opportunity: Why outcome-based AI solutions are targeting the $35 trillion labor market instead of just software budgets and how that changes everything for startups and investors.

Voice AI’s healthcare breakthrough: How voice agents are finally solving the operational bottlenecks in patient scheduling and communication, driving 24/7 availability with better NPS than human operators.

Why 95% of GenAI projects still fail: The technical and infrastructure gaps that prevent most AI initiatives from making it to production and what’s needed to fix that in 2026.

The new technical risk era: After years of focusing purely on market risk, VCs are back to evaluating deep technical challenges in agentic systems, browser automation, and continuous learning loops.

The exceptionalism filter: How early-stage investors are separating signal from noise when everyone can spin up an AI startup and why founder insights and lived experience matter more than ever.

Timestamped Highlights

00:31 – What Felicis invests in and the types of AI companies dominating their portfolio right now

02:58 – Why healthcare tech is finally ready for its AI moment after years of long sales cycles and unclear ROI

08:15 – How outcome-based pricing is changing the VC evaluation playbook and unlocking 10x larger TAMs

13:26 – The mythical one-person billion-dollar company: Is it real, and how would investors even spot it?

17:18 – Voice AI as the gateway for enterprise adoption and why this modality is different from Siri and Alexa

20:08 – Democratizing AI: What ChatGPT did for consumers and what needs to happen for enterprise builders

One Thing Worth Remembering

“These technologies can price towards the labor replacement markets, which is about 10x the size of the software market itself. The ROI right now is so tangible that it is a time to invest.”

Subscribe and Stay in the Loop

If this episode gave you a new angle on where AI is actually delivering value, share it with a founder or investor who needs to hear it. Subscribe so you don’t miss the next conversation, and drop a comment if there’s a topic or guest you want us to tackle next.

In this episode, Amir sits down with Taofeek Rabiu, VP of Engineering at Etsy, to unpack a distinction that most organizations miss: being a people leader is not the same as being a people manager.

If you have ever wondered why some teams thrive under pressure while others crumble, or why trust feels so hard to build in engineering orgs, this conversation has answers. Taofeek shares how leadership is not reserved for those with a manager title, why vulnerability is a strategic advantage, and how to spot the early warning signs of poor leadership before they drag down performance.

What You’ll Learn

Leadership exists at every level, not just in management roles. Individual contributors who mentor, influence, and model the right behaviors are leaders too — and organizations need to recognize and reward that.

Trust is built through action, not talk. It grows when leaders show vulnerability, stay transparent about their thinking, and follow through on commitments. When you stop acting on what you hear, you break trust.

Poor leadership has a smell. Teams that avoid hard conversations, struggle to navigate change, or fail to ramp new hires are showing symptoms of leadership gaps, not process problems.

Feedback is about helping people see, not telling them what to do. The best leaders use curiosity to guide others toward realization and self-awareness.

Effective leaders make high signal, low frequency decisions. The goal is not to make a thousand calls a day but to gather diverse perspectives and make the few decisions that truly move the team forward.

Timestamped Highlights

01:42 – Taofeek breaks down the difference between managing people (reviews, org charts, timesheets) and leading people (building trust, showing care, creating psychological safety).

09:04 – What happens when managers focus only on mechanics. Taofeek describes the smells of poor leadership and how they surface in teams that can’t handle change.

13:18 – How to give feedback when someone is not showing up as a leader. Taofeek explains his approach: start with curiosity, triangulate with skip levels, and guide people to their own realizations.

17:47 – Who is responsible for building trust. Taofeek shares why it is on leaders to create the conditions, not on reports to earn it.

22:04 – The moment Simon Sinek told Taofeek to stop saying people managers and start saying people leaders — and how that small shift in language changed his approach to leadership.

24:29 – What feedback a VP of Engineering actually values. Taofeek shares how he uncovers blind spots and the kind of input that helps him grow.

Words That Stuck

“The team doesn’t trust you. You’re not providing a psychologically safe environment in which the team feels like they can course correct and flag things that they believe will lead to poor outcomes.”

If This Resonates, Here’s What to Do

Take one insight from this episode and put it into practice this week. Maybe it’s being more open in your next one-on-one, checking your follow-through, or asking your team a question you have been avoiding. Then share this episode with someone navigating the manager-to-leader transition. Subscribe to The Tech Trek for more conversations that help you grow as a leader, and connect with Taofeek on LinkedIn to keep the dialogue going.

Ion Feldman, CTO at Rightway, has learned to love one thing about scaling a company from a kitchen table to nearly 1,000 employees: his job completely changes every six months. In this episode, Ion shares what it means to lead engineering when the role refuses to stay still—from writing code in the early days to building product, security, and data teams, and now shaping AI infrastructure. He explains how to stay hands-on without micromanaging, why he deliberately works himself out of roles by hiring people better than him, and how to preserve startup urgency inside a heavily regulated industry. If you’ve ever wondered how CTOs balance technical depth with business strategy while keeping their team fast and focused, this conversation delivers.

Key Takeaways

Treat change as part of the job.

Ion’s leadership mindset centers on adapting to wherever the company needs him most—product, security, data, or AI. He views change as an opportunity to grow, not a disruption to avoid.

Hire yourself out of the role.

He dives deep into an area, builds it from scratch, then brings in experts who can take it to the next level. Once the right leadership is in place, he steps back completely and lets them own it.

Hands-on time creates credibility.

Ion makes sure every leader spends time building. Each quarter, his team takes a week off from meetings and Slack to focus on creating something new. It keeps them close to the work and sharp as technical leaders.

AI adoption needs clarity and focus.

Rightway avoids vague “use AI” goals by targeting clear use cases like unit test generation and onboarding to codebases. Sharing examples and results drives faster adoption than leaving teams to figure it out alone.

Fail fast and move forward.

Ion builds space for experimentation but expects quick recognition of failure. The goal is not to avoid mistakes but to learn, pivot, and evolve faster.

Timestamped Highlights

[02:10] The zero to one mindset – Why Ion thrives on constant reinvention and the satisfaction of building new functions from the ground up.

[06:41] Three pillars of AI strategy – How Rightway is transforming work through AI enablement, applied projects, and bold experiments.

[08:26] Delegating by design – How going deep before handing off creates clarity and trust across teams.

[15:42] Skills that matter later – Ion reflects on learning public speaking and business fluency after years of technical focus.

[17:48] Creating space for risk – How to give your team agency to take on big challenges and fail fast without fear.

[21:22] Preparing successors – Why the best leaders hire people who will replace them and rethink everything they built.

What Stuck With Us

"I don't know, maybe I just get bored easily. I think a lot of people could view it as a burden and they want to stay in their lane of expertise, but I see it as an opportunity to learn and change things up."

Pro Tips for Tech Leaders

Take a week each quarter to build something with zero meetings or Slack. It reconnects you and your team with what you actually love about engineering.

Wait to hire senior leadership until the need is undeniable. The role becomes meaningful, and you’ll attract higher caliber talent.

Give your engineers specific AI examples and let them experiment from there. Adoption follows clarity, not mandates.