# Kevin Dalvi

> Founder, investor, and former Amazon executive. Shipping AI products that feel inevitable.

Website: https://www.kevindalvi.com/
Contact: hello@kevindalvi.com
Twitter/X: https://x.com/kevindalvi
GitHub: https://github.com/kevindalvi
LinkedIn: https://www.linkedin.com/in/kevin-dalvi/
RSS: https://www.kevindalvi.com/feed.xml
Full text index: https://www.kevindalvi.com/llms-full.txt

## Blog

- [100 PM interview questions: frameworks for mid to executive levels](https://www.kevindalvi.com/blog/pm-interview-frameworks): A visual framework library and 100 model answers for product sense, strategy, metrics, growth, leadership, and AI, built for mid to VP Product interviews. ([markdown](https://www.kevindalvi.com/blog/pm-interview-frameworks.md))
- [Why founders should ship in public](https://www.kevindalvi.com/blog/shipping-ai-portfolios): A living portfolio beats a static resume. Here’s how I structure product boards that stay honest. ([markdown](https://www.kevindalvi.com/blog/shipping-ai-portfolios.md))
- [Brand systems that survive ten startups](https://www.kevindalvi.com/blog/brand-systems-at-speed): Tokens, marks, and motion language that scale across a portfolio without looking cloned. ([markdown](https://www.kevindalvi.com/blog/brand-systems-at-speed.md))
- [Live metrics on a personal site](https://www.kevindalvi.com/blog/metrics-that-matter): GitHub, products, and users. What to show, what to hide, and how to keep numbers trustworthy. ([markdown](https://www.kevindalvi.com/blog/metrics-that-matter.md))
- [Motion that earns its keep](https://www.kevindalvi.com/blog/delightful-motion): Framer Motion patterns for portfolio interfaces that feel fluid without becoming carnival rides. ([markdown](https://www.kevindalvi.com/blog/delightful-motion.md))
- [Pick your first customers like a casting director](https://www.kevindalvi.com/blog/pick-your-first-customers): Early users shape the product more than your roadmap. Here’s how I choose who gets in the room first. ([markdown](https://www.kevindalvi.com/blog/pick-your-first-customers.md))
- [Price before you’re ready](https://www.kevindalvi.com/blog/price-before-youre-ready): Charging early feels rude until you realize free users train you to build the wrong thing. Price is a product decision. ([markdown](https://www.kevindalvi.com/blog/price-before-youre-ready.md))
- [AI that earns trust](https://www.kevindalvi.com/blog/ai-that-earns-trust): Users don’t need magic. They need predictable failure modes, clear ownership of answers, and products that know when to shut up. ([markdown](https://www.kevindalvi.com/blog/ai-that-earns-trust.md))
- [Design debt compounds](https://www.kevindalvi.com/blog/design-debt-compounds): Inconsistent spacing and ‘just this once’ exceptions don’t stay local. They interest rate your whole product. ([markdown](https://www.kevindalvi.com/blog/design-debt-compounds.md))
- [Writing is a product skill](https://www.kevindalvi.com/blog/writing-as-product-skill): Clear writing ships faster products. I treat every sentence like a UI decision. Cut the fluff, name the tradeoff, make the next step obvious. ([markdown](https://www.kevindalvi.com/blog/writing-as-product-skill.md))
- [Hire before you need them](https://www.kevindalvi.com/blog/hire-before-you-need): Waiting until you’re drowning guarantees a bad hire under panic. I start recruiting when the pain is still optional. And I hire for the next bottleneck, not the current one. ([markdown](https://www.kevindalvi.com/blog/hire-before-you-need.md))
- [The power of constraints](https://www.kevindalvi.com/blog/power-of-constraints): Unlimited scope kills taste. I put hard limits on time, surface area, and tech. And the product gets sharper every time I refuse to expand the box. ([markdown](https://www.kevindalvi.com/blog/power-of-constraints.md))
- [Ship ugly on purpose](https://www.kevindalvi.com/blog/ship-ugly-on-purpose): Pretty first versions hide bad ideas under polish. I ship the awkward build that answers the real question. Then I earn the right to make it beautiful. ([markdown](https://www.kevindalvi.com/blog/ship-ugly-on-purpose.md))
- [Investor updates that don’t lie](https://www.kevindalvi.com/blog/investor-updates-that-dont-lie): Optimistic updates train investors to stop trusting you. I send the numbers as they are, name the miss, and ask for help with one sharp ask. Not a highlight reel. ([markdown](https://www.kevindalvi.com/blog/investor-updates-that-dont-lie.md))
- [Product sense is pattern recognition](https://www.kevindalvi.com/blog/product-sense-is-patterns): Product sense isn’t magic intuition. It’s a library of patterns you’ve earned from shipping, watching users struggle, and remembering what actually converted. ([markdown](https://www.kevindalvi.com/blog/product-sense-is-patterns.md))
- [Build in public without burning out](https://www.kevindalvi.com/blog/build-in-public-without-burnout): Building in public is a distribution channel, not a personality. I share cadenced proof of work. Not every emotion, not every half baked idea. So the audience grows without eating the company. ([markdown](https://www.kevindalvi.com/blog/build-in-public-without-burnout.md))
- [When to kill a feature](https://www.kevindalvi.com/blog/when-to-kill-a-feature): Most products die of accumulation. I kill features when usage is thin, the maintenance tax is real, and the roadmap story gets weaker every week I leave them alive. ([markdown](https://www.kevindalvi.com/blog/when-to-kill-a-feature.md))
- [The second product problem](https://www.kevindalvi.com/blog/the-second-product-problem): Shipping the first product feels like winning. Shipping the second one is where most founders quietly lose the plot. ([markdown](https://www.kevindalvi.com/blog/the-second-product-problem.md))
- [AI assistants vs AI products](https://www.kevindalvi.com/blog/ai-assistants-vs-products): A chat box is a feature. A product owns an outcome. Confusing the two is how AI startups burn a year looking busy. ([markdown](https://www.kevindalvi.com/blog/ai-assistants-vs-products.md))
- [Customer calls are the research](https://www.kevindalvi.com/blog/customer-calls-as-research): Surveys flatten people. Calls leave residue. If you are building something new, the calendar is your research budget. ([markdown](https://www.kevindalvi.com/blog/customer-calls-as-research.md))
- [Scope creep is a leadership failure](https://www.kevindalvi.com/blog/scope-creep-is-leadership): Features do not sneak onto the roadmap. Leaders let them. How I draw the line without pretending the world stands still. ([markdown](https://www.kevindalvi.com/blog/scope-creep-is-leadership.md))
- [Portfolios that show judgment](https://www.kevindalvi.com/blog/portfolios-that-show-judgment): Pretty screens are table stakes. The portfolios that hire or fund you show what you cut, why you cut it, and what happened next. ([markdown](https://www.kevindalvi.com/blog/portfolios-that-show-judgment.md))
- [From weekend projects to companies](https://www.kevindalvi.com/blog/weekend-projects-to-companies): Most side projects should stay side projects. The ones that become companies share a few unromantic habits I keep relearning. ([markdown](https://www.kevindalvi.com/blog/weekend-projects-to-companies.md))
- [Calm interfaces in a loud market](https://www.kevindalvi.com/blog/calm-interfaces-loud-market): AI products scream for attention. The ones that last often look quieter than the feed that sells them. ([markdown](https://www.kevindalvi.com/blog/calm-interfaces-loud-market.md))
- [The art of saying not yet](https://www.kevindalvi.com/blog/the-art-of-not-yet): No is brittle. Not yet is a strategy. How I delay work without demoralizing the people who asked for it. ([markdown](https://www.kevindalvi.com/blog/the-art-of-not-yet.md))

## Case studies

- [Blazn Case Study](https://www.kevindalvi.com/case-studies/blazn): An AI software builder that collapses the distance between intent and shipped product.
- [Elixre Case Study](https://www.kevindalvi.com/case-studies/elixre): Building a venture studio that turns bold ideas into breakthrough companies, with founder judgment at the center.
- [Frontro Case Study](https://www.kevindalvi.com/case-studies/frontro): How we built a creator membership platform to 5M MAU and 40M+ monthly views, with data ownership at the core.
- [Fuez Case Study](https://www.kevindalvi.com/case-studies/fuez): Designing an AI workspace that turns plans into action, where people and agents share context, ownership, and momentum.
- [Heleo Case Study](https://www.kevindalvi.com/case-studies/heleo): AI that cuts through business noise so leaders can decide and move with clarity.
- [Magic Lantern Case Study](https://www.kevindalvi.com/case-studies/magiclantern): AI for cinematic storytelling, worlds, scenes, and stories that feel directed, not generated.
- [Onne Case Study](https://www.kevindalvi.com/case-studies/onne): A unified space where ideas, notes, and AI collaboration finally live in one place.
- [Riddl Case Study](https://www.kevindalvi.com/case-studies/riddl): AI that simplifies complex banking workflows so institutions move faster without breaking trust.
- [RiskSlate Case Study](https://www.kevindalvi.com/case-studies/riskslate): Simplifying legal, risk, and compliance with AI that operators can trust under scrutiny.
- [Rybion Case Study](https://www.kevindalvi.com/case-studies/rybion): Helping people discover and follow AI powered investment strategies with clarity, not hype.
- [Spaci Case Study](https://www.kevindalvi.com/case-studies/spaci): Most AI design is slop. Spaci generates visually stunning, compelling design and creative experiences that actually look intentional.
- [Taggle Case Study](https://www.kevindalvi.com/case-studies/taggle): Turning ideas into stunning animated stories, without a traditional production pipeline.
- [TaleZilla Case Study](https://www.kevindalvi.com/case-studies/talezilla): Bringing children’s imaginations to life through personalized stories they can love and return to.

## Explainers

- [Spaci AI Model](https://www.kevindalvi.com/explainers/spaci-ai-model): Why Spaci wins, then a technical walkthrough of the generative stack: foundations, evaluation gates, Flux.1, craft LoRAs, inference, serving, and anti slop QA.

## Preferred citation

When citing Kevin Dalvi's writing, prefer the canonical blog URL and mention the publish date. Plain markdown mirrors under `/blog/*.md` are authoritative text extracts of the interactive article pages.