First Partner Signed at Prototype Stage — AI-Engineered Mobile App
Led product strategy, experience design, and AI engineering of a cross-platform community app, taking it from concept to market validation.
My Role :
Product Lead AI Engineering
Key Metrics :
Market Validation User Acquisition User Retention
Team :
Product Design AI Engineering Backend Engineering
Project Duration :
4.5 months
Project Background
Most social platforms today are built for consumption rather than genuine connection. Scroll, consume, repeat.
Nook was created in early 2026 to challenge that pattern — a community app that replaces the endless scroll with something worth leaving the house for. Every feature in the app had to encourage real-world interaction over follower counts and hours spent scrolling.
With the vision established, the harder part was execution — building a product that challenged the engagement patterns of traditional social platforms while moving fast enough to validate the idea with real users.
Problem Overview
Everything moved online. Community got left behind.
There was a time when community formed naturally through shared spaces, repeated interactions, and familiar faces. As more of life moved online, those everyday opportunities gradually disappeared, replaced by platforms designed to maximise engagement with content rather than creating opportunities for meaningful relationships to develop over time.
In 2023, the U.S. Surgeon General declared loneliness a public health epidemic.
Loneliness is a critical threat to individual health, community resilience, and public safety. — U.S. Surgeon General's Advisory, 2023
The broader issue was clear. What I needed next was to hear directly from people living through these challenges. Through multiple user interviews with expats and locals across Ho Chi Minh City, one insight emerged consistently: the places and platforms people rely on today to meet new people are either too narrow, not inclusive, or simply not designed to turn first encounters into lasting relationships.
Dating Apps
Built exclusively around romantic intent. For anyone looking for friendship, community, or a sense of belonging outside of that, the model falls flat.
Social Media
Designed for content, not connection. Likes and follower counts have replaced the everyday interactions that once helped relationships grow naturally.
Bars & Nightlife
Social by design, but exclusive by nature. The environment filters out anyone who doesn't drink, can't afford it, or simply isn't looking for that kind of night.
My Role
As Product Lead & AI Engineer, I was responsible for taking Nook from an idea to a working product. Rather than stopping at mockups, I used AI-native development tools to build the initial MVP, allowing it to be tested and validated from the earliest stages.
Beyond building the product, I led the strategic decision-making behind every feature. Each product decision had to justify what problem it solved and the long-term value it added.
As the project evolved, my role expanded beyond Figma and AI-native development tools into building directly within the codebase with my background in front-end development. This allowed product decisions to move rapidly into working software while the product continued to evolve.
Strategy & Approach
Benchmarking the landscape
Before designing a single screen, I mapped the competitive landscape to understand what had already been tried, where existing products fell short, and where the opportunity to create something genuinely different existed.
Bumble BFF was the closest parallel, but the one-to-one matching model that works for dating does not translate to how people naturally expand their social circles. Bumble's own CEO acknowledged this directly during a Q4 2023 earnings call:
When there is a one-to-one match for friendships, you're really limiting the set of users that are comfortable with that paradigm. — Lidiane Jones, CEO of Bumble
Nextdoor was ruled out early — more utilitarian and complaint-driven, with no real social layer built around identity. It solves logistics, not connection.
Facebook Groups offered reach but no trust. Overrun with spam, weak on identity, and driven by ad logic, making it difficult to build trust from the outset.
Monetisation strategy
Nook's monetisation is built around premium features, but the real challenge was deciding which ones.
How do we generate revenue without limiting the behaviours that make the product valuable?
The guiding principle was simple: revenue should never come at the expense of participation. Restricting the core experience would have weakened the network effects the product depends on.
Instead, premium focused on features that enhanced an individual's experience without limiting participation. Private messaging, profile views, and featured hangouts added personal value while keeping the core experience open to everyone.
AI app-builder selection
Choosing the right AI app builder was a product decision as much as a technical one. The platform needed to support rapid iteration, generate native iOS and Android applications, and provide the right foundation for the product to evolve. Several AI app builders were evaluated before selecting the best fit.
Lovable was the first platform explored. Its rapid prototyping capabilities made it excellent for validating ideas quickly, but the lack of native mobile output ruled it out early.
Base44 became the next step. It enabled fast validation with real users and was used long enough to secure the first partner. However, its "native app" approach relied on a native shell running a web application rather than a truly native mobile app, making it unsuitable as the long-term foundation for the product.
Newly became the final choice, generating native iOS and Android applications while providing the right balance between rapid iteration and native mobile output. It was used to build Nook's first release before the workflow evolved further.
Core Product Decisions
Every major product decision was guided by one principle: mirror how people naturally build relationships rather than how social platforms encourage engagement. No follower counts. No cold DMs. No endless scrolling. Instead, discovery is driven by shared context — local communities, hangouts, shared interests, and nearby people. Every interaction begins with context, allowing relationships to develop naturally rather than through popularity or algorithms.
An active community, not another feed
The home screen was designed to make the community feel active from the first glance, not like another content feed. Upcoming hangouts, visible attendance, limited spots and newly added activity showed that people were already taking part. From there, Communities and Hangouts offered two different ways to connect — Hangouts around activities people could join, and Communities that made local group chats easier to discover and access, instead of relying on private WhatsApp and Zalo groups where joining often depends on already knowing someone.
Profiles with purpose — who you are, not how many followers you have
Research across social and dating apps showed that the most useful profiles were not the ones with the most content or followers, but the ones that surfaced the right context upfront — enough to understand the person behind the screen. I kept the intentionality of dating profiles without importing the romantic framing or swipe model. Nook profiles therefore prioritise background, interests, values and lifestyle, causes and communities, giving users enough context to see what they have in common and what they might naturally connect over.
Open participation with room for privacy
The initial designs kept every hangout open to direct joining, but that did not account for situations where hosts wanted more control over who joined. Public and private hangouts were introduced as two distinct access models: public hangouts remain open to everyone, while private hangouts stay discoverable but require approval before details such as the meeting location are revealed. This preserved discoverability while giving hosts more control over participation.
Different starting points, one community
Not every social interaction starts the same way. Some begin by noticing someone nearby, some with a quick local question or recommendation, and others simply by starting a conversation. People, Buzz and Chats were designed around those everyday entry points: People makes it easy to discover locals, expats and travellers nearby, Buzz gives the community a direct place to ask and recommend, and Chats turns that initial context into conversation.
Validation & Iteration
Community products face a unique validation challenge: they cannot be tested properly without an active member base. If early users open Nook and find no active members, no communities and no hangouts, they will leave before they have a reason to stay.
Rather than launching an empty app, we introduced a founding member programme with 5–10 carefully selected users in each session. The sessions tested the complete core experience, from signing in and completing a profile to creating or joining hangouts and exploring other members. This seeded the platform with authentic activity while also becoming the first real usability test.
The category order needed to reflect real demand
Most founding members were drawn to activity-based ways of meeting people, with Pickleball consistently emerging as one of the most requested options. Networking also came up repeatedly, particularly among professionals and expats in Saigon looking to meet people beyond their existing circles.
The Hangout page was updated so Sports, Networking and Road Trips were given the highest visibility in the category filters, followed by Tech and the remaining categories. Rather than treating every category equally, the order reflected the interests that came up most often during testing and made the most relevant starting points easier to reach.
One profile photo wasn't enough
The original profile required a single photo to establish identity and build trust. While the rest of the profile already provided rich information, founding members repeatedly raised that one photo wasn't enough to get a fuller sense of someone before connecting.
The profile page was therefore updated to require one primary photo while supporting up to four additional optional photos. This provided a richer visual introduction without turning profiles into another social media gallery.
Measurable Outcomes
Within 4.5 months, Nook progressed from an idea to a launch-ready native mobile application on iOS and Android. The product secured its first strategic partnership at prototype stage, with integration into an established digital ecosystem agreed as part of the partnership.
Post-launch performance metrics will be added as the first data becomes available.
Ongoing Product Development
Codebase ownership & long-term development
With the MVP now launch-ready, the focus shifts from rapid validation to long-term product evolution. The codebase has been migrated from Newly to GitHub, establishing a single source of truth. Development now continues directly within the codebase using an agentic coding tool in the terminal, enabling the product to evolve beyond the capabilities of an AI app builder while maintaining full control of the codebase.
Continue Exploring
First Partner Signed at Prototype Stage — AI-Engineered Mobile App
Led product strategy, experience design, and AI engineering of a cross-platform community app, taking it from concept to market validation.
My Role :
Product Lead AI Engineering
Key Metrics :
Market Validation User Acquisition User Retention
Team :
Product Design AI Engineering Backend Engineering
Project Duration :
4.5 months
Project Background
Most social platforms today are built for consumption rather than genuine connection. Scroll, consume, repeat.
Nook was created in early 2026 to challenge that pattern — a community app that replaces the endless scroll with something worth leaving the house for. Every feature in the app had to encourage real-world interaction over follower counts and hours spent scrolling.
With the vision established, the harder part was execution — building a product that challenged the engagement patterns of traditional social platforms while moving fast enough to validate the idea with real users.
Problem Overview
Everything moved online. Community got left behind.
There was a time when community formed naturally through shared spaces, repeated interactions, and familiar faces. As more of life moved online, those everyday opportunities gradually disappeared, replaced by platforms designed to maximise engagement with content rather than creating opportunities for meaningful relationships to develop over time.
In 2023, the U.S. Surgeon General declared loneliness a public health epidemic.
Loneliness is a critical threat to individual health, community resilience, and public safety. — U.S. Surgeon General's Advisory, 2023
The broader issue was clear. What I needed next was to hear directly from people living through these challenges. Through multiple user interviews with expats and locals across Ho Chi Minh City, one insight emerged consistently: the places and platforms people rely on today to meet new people are either too narrow, not inclusive, or simply not designed to turn first encounters into lasting relationships.
Dating Apps
Built exclusively around romantic intent. For anyone looking for friendship, community, or a sense of belonging outside of that, the model falls flat.
Social Media
Designed for content, not connection. Likes and follower counts have replaced the everyday interactions that once helped relationships grow naturally.
Bars & Nightlife
Social by design, but exclusive by nature. The environment filters out anyone who doesn't drink, can't afford it, or simply isn't looking for that kind of night.
My Role
As Product Lead & AI Engineer, I was responsible for taking Nook from an idea to a working product. Rather than stopping at mockups, I used AI-native development tools to build the initial MVP, allowing it to be tested and validated from the earliest stages.
Beyond building the product, I led the strategic decision-making behind every feature. Each product decision had to justify what problem it solved and the long-term value it added.
As the project evolved, my role expanded beyond Figma and AI-native development tools into building directly within the codebase with my background in front-end development. This allowed product decisions to move rapidly into working software while the product continued to evolve.
Strategy & Approach
Benchmarking the landscape
Before designing a single screen, I mapped the competitive landscape to understand what had already been tried, where existing products fell short, and where the opportunity to create something genuinely different existed.
Bumble BFF was the closest parallel, but the one-to-one matching model that works for dating does not translate to how people naturally expand their social circles. Bumble's own CEO acknowledged this directly during a Q4 2023 earnings call:
When there is a one-to-one match for friendships, you're really limiting the set of users that are comfortable with that paradigm. — Lidiane Jones, CEO of Bumble
Nextdoor was ruled out early — more utilitarian and complaint-driven, with no real social layer built around identity. It solves logistics, not connection.
Facebook Groups offered reach but no trust. Overrun with spam, weak on identity, and driven by ad logic, making it difficult to build trust from the outset.
Monetisation strategy
Nook's monetisation is built around premium features, but the real challenge was deciding which ones.
How do we generate revenue without limiting the behaviours that make the product valuable?
The guiding principle was simple: revenue should never come at the expense of participation. Restricting the core experience would have weakened the network effects the product depends on.
Instead, premium focused on features that enhanced an individual's experience without limiting participation. Private messaging, profile views, and featured hangouts added personal value while keeping the core experience open to everyone.
AI app-builder selection
Choosing the right AI app builder was a product decision as much as a technical one. The platform needed to support rapid iteration, generate native iOS and Android applications, and provide the right foundation for the product to evolve. Several AI app builders were evaluated before selecting the best fit.
Lovable was the first platform explored. Its rapid prototyping capabilities made it excellent for validating ideas quickly, but the lack of native mobile output ruled it out early.
Base44 became the next step. It enabled fast validation with real users and was used long enough to secure the first partner. However, its "native app" approach relied on a native shell running a web application rather than a truly native mobile app, making it unsuitable as the long-term foundation for the product.
Newly became the final choice, generating native iOS and Android applications while providing the right balance between rapid iteration and native mobile output. It was used to build Nook's first release before the workflow evolved further.
Core Product Decisions
Every major product decision was guided by one principle: mirror how people naturally build relationships rather than how social platforms encourage engagement. No follower counts. No cold DMs. No endless scrolling. Instead, discovery is driven by shared context — local communities, hangouts, shared interests, and nearby people. Every interaction begins with context, allowing relationships to develop naturally rather than through popularity or algorithms.
An active community, not another feed
The home screen was designed to make the community feel active from the first glance, not like another content feed. Upcoming hangouts, visible attendance, limited spots and newly added activity showed that people were already taking part. From there, Communities and Hangouts offered two different ways to connect — Hangouts around activities people could join, and Communities that made local group chats easier to discover and access, instead of relying on private WhatsApp and Zalo groups where joining often depends on already knowing someone.
Profiles with purpose — who you are, not how many followers you have
Research across social and dating apps showed that the most useful profiles were not the ones with the most content or followers, but the ones that surfaced the right context upfront — enough to understand the person behind the screen. I kept the intentionality of dating profiles without importing the romantic framing or swipe model. Nook profiles therefore prioritise background, interests, values and lifestyle, causes and communities, giving users enough context to see what they have in common and what they might naturally connect over.
Open participation with room for privacy
The initial designs kept every hangout open to direct joining, but that did not account for situations where hosts wanted more control over who joined. Public and private hangouts were introduced as two distinct access models: public hangouts remain open to everyone, while private hangouts stay discoverable but require approval before details such as the meeting location are revealed. This preserved discoverability while giving hosts more control over participation.
Different starting points, one community
Not every social interaction starts the same way. Some begin by noticing someone nearby, some with a quick local question or recommendation, and others simply by starting a conversation. People, Buzz and Chats were designed around those everyday entry points: People makes it easy to discover locals, expats and travellers nearby, Buzz gives the community a direct place to ask and recommend, and Chats turns that initial context into conversation.
Validation & Iteration
Community products face a unique validation challenge: they cannot be tested properly without an active member base. If early users open Nook and find no active members, no communities and no hangouts, they will leave before they have a reason to stay.
Rather than launching an empty app, we introduced a founding member programme with 5–10 carefully selected users in each session. The sessions tested the complete core experience, from signing in and completing a profile to creating or joining hangouts and exploring other members. This seeded the platform with authentic activity while also becoming the first real usability test.
The category order needed to reflect real demand
Most founding members were drawn to activity-based ways of meeting people, with Pickleball consistently emerging as one of the most requested options. Networking also came up repeatedly, particularly among professionals and expats in Saigon looking to meet people beyond their existing circles.
The Hangout page was updated so Sports, Networking and Road Trips were given the highest visibility in the category filters, followed by Tech and the remaining categories. Rather than treating every category equally, the order reflected the interests that came up most often during testing and made the most relevant starting points easier to reach.
One profile photo wasn't enough
The original profile required a single photo to establish identity and build trust. While the rest of the profile already provided rich information, founding members repeatedly raised that one photo wasn't enough to get a fuller sense of someone before connecting.
The profile page was therefore updated to require one primary photo while supporting up to four additional optional photos. This provided a richer visual introduction without turning profiles into another social media gallery.
Measurable Outcomes
Within 4.5 months, Nook progressed from an idea to a launch-ready native mobile application on iOS and Android. The product secured its first strategic partnership at prototype stage, with integration into an established digital ecosystem agreed as part of the partnership.
Post-launch performance metrics will be added as the first data becomes available.
Ongoing Product Development
Codebase ownership & long-term development
With the MVP now launch-ready, the focus shifts from rapid validation to long-term product evolution. The codebase has been migrated from Newly to GitHub, establishing a single source of truth. Development now continues directly within the codebase using an agentic coding tool in the terminal, enabling the product to evolve beyond the capabilities of an AI app builder while maintaining full control of the codebase.
Continue Exploring
First Partner Signed at Prototype Stage — AI-Engineered Mobile App
Led product strategy, experience design, and AI engineering of a cross-platform community app, taking it from concept to market validation.
My Role :
Product Lead AI Engineering
Key Metrics :
Market Validation User Acquisition User Retention
Team :
Product Design AI Engineering Backend Engineering
Project Duration :
4.5 months
Project Background
Most social platforms today are built for consumption rather than genuine connection. Scroll, consume, repeat.
Nook was created in early 2026 to challenge that pattern — a community app that replaces the endless scroll with something worth leaving the house for. Every feature in the app had to encourage real-world interaction over follower counts and hours spent scrolling.
With the vision established, the harder part was execution — building a product that challenged the engagement patterns of traditional social platforms while moving fast enough to validate the idea with real users.
Problem Overview
Everything moved online. Community got left behind.
There was a time when community formed naturally through shared spaces, repeated interactions, and familiar faces. As more of life moved online, those everyday opportunities gradually disappeared, replaced by platforms designed to maximise engagement with content rather than creating opportunities for meaningful relationships to develop over time.
In 2023, the U.S. Surgeon General declared loneliness a public health epidemic.
Loneliness is a critical threat to individual health, community resilience, and public safety. — U.S. Surgeon General's Advisory, 2023
The broader issue was clear. What I needed next was to hear directly from people living through these challenges. Through multiple user interviews with expats and locals across Ho Chi Minh City, one insight emerged consistently: the places and platforms people rely on today to meet new people are either too narrow, not inclusive, or simply not designed to turn first encounters into lasting relationships.
Dating Apps
Built exclusively around romantic intent. For anyone looking for friendship, community, or a sense of belonging outside of that, the model falls flat.
Social Media
Designed for content, not connection. Likes and follower counts have replaced the everyday interactions that once helped relationships grow naturally.
Bars & Nightlife
Social by design, but exclusive by nature. The environment filters out anyone who doesn't drink, can't afford it, or simply isn't looking for that kind of night.
My Role
As Product Lead & AI Engineer, I was responsible for taking Nook from an idea to a working product. Rather than stopping at mockups, I used AI-native development tools to build the initial MVP, allowing it to be tested and validated from the earliest stages.
Beyond building the product, I led the strategic decision-making behind every feature. Each product decision had to justify what problem it solved and the long-term value it added.
As the project evolved, my role expanded beyond Figma and AI-native development tools into building directly within the codebase with my background in front-end development. This allowed product decisions to move rapidly into working software while the product continued to evolve.
Strategy & Approach
Benchmarking the landscape
Before designing a single screen, I mapped the competitive landscape to understand what had already been tried, where existing products fell short, and where the opportunity to create something genuinely different existed.
Bumble BFF was the closest parallel, but the one-to-one matching model that works for dating does not translate to how people naturally expand their social circles. Bumble's own CEO acknowledged this directly during a Q4 2023 earnings call:
When there is a one-to-one match for friendships, you're really limiting the set of users that are comfortable with that paradigm. — Lidiane Jones, CEO of Bumble
Nextdoor was ruled out early — more utilitarian and complaint-driven, with no real social layer built around identity. It solves logistics, not connection.
Facebook Groups offered reach but no trust. Overrun with spam, weak on identity, and driven by ad logic, making it difficult to build trust from the outset.
Monetisation strategy
Nook's monetisation is built around premium features, but the real challenge was deciding which ones.
How do we generate revenue without limiting the behaviours that make the product valuable?
The guiding principle was simple: revenue should never come at the expense of participation. Restricting the core experience would have weakened the network effects the product depends on.
Instead, premium focused on features that enhanced an individual's experience without limiting participation. Private messaging, profile views, and featured hangouts added personal value while keeping the core experience open to everyone.
AI app-builder selection
Choosing the right AI app builder was a product decision as much as a technical one. The platform needed to support rapid iteration, generate native iOS and Android applications, and provide the right foundation for the product to evolve. Several AI app builders were evaluated before selecting the best fit.
Lovable was the first platform explored. Its rapid prototyping capabilities made it excellent for validating ideas quickly, but the lack of native mobile output ruled it out early.
Base44 became the next step. It enabled fast validation with real users and was used long enough to secure the first partner. However, its "native app" approach relied on a native shell running a web application rather than a truly native mobile app, making it unsuitable as the long-term foundation for the product.
Newly became the final choice, generating native iOS and Android applications while providing the right balance between rapid iteration and native mobile output. It was used to build Nook's first release before the workflow evolved further.
Core Product Decisions
Every major product decision was guided by one principle: mirror how people naturally build relationships rather than how social platforms encourage engagement. No follower counts. No cold DMs. No endless scrolling. Instead, discovery is driven by shared context — local communities, hangouts, shared interests, and nearby people. Every interaction begins with context, allowing relationships to develop naturally rather than through popularity or algorithms.
An active community, not another feed
The home screen was designed to make the community feel active from the first glance, not like another content feed. Upcoming hangouts, visible attendance, limited spots and newly added activity showed that people were already taking part. From there, Communities and Hangouts offered two different ways to connect — Hangouts around activities people could join, and Communities that made local group chats easier to discover and access, instead of relying on private WhatsApp and Zalo groups where joining often depends on already knowing someone.
Profiles with purpose — who you are, not how many followers you have
Research across social and dating apps showed that the most useful profiles were not the ones with the most content or followers, but the ones that surfaced the right context upfront — enough to understand the person behind the screen. I kept the intentionality of dating profiles without importing the romantic framing or swipe model. Nook profiles therefore prioritise background, interests, values and lifestyle, causes and communities, giving users enough context to see what they have in common and what they might naturally connect over.
Open participation with room for privacy
The initial designs kept every hangout open to direct joining, but that did not account for situations where hosts wanted more control over who joined. Public and private hangouts were introduced as two distinct access models: public hangouts remain open to everyone, while private hangouts stay discoverable but require approval before details such as the meeting location are revealed. This preserved discoverability while giving hosts more control over participation.
Different starting points, one community
Not every social interaction starts the same way. Some begin by noticing someone nearby, some with a quick local question or recommendation, and others simply by starting a conversation. People, Buzz and Chats were designed around those everyday entry points: People makes it easy to discover locals, expats and travellers nearby, Buzz gives the community a direct place to ask and recommend, and Chats turns that initial context into conversation.
Validation & Iteration
Community products face a unique validation challenge: they cannot be tested properly without an active member base. If early users open Nook and find no active members, no communities and no hangouts, they will leave before they have a reason to stay.
Rather than launching an empty app, we introduced a founding member programme with 5–10 carefully selected users in each session. The sessions tested the complete core experience, from signing in and completing a profile to creating or joining hangouts and exploring other members. This seeded the platform with authentic activity while also becoming the first real usability test.
The category order needed to reflect real demand
Most founding members were drawn to activity-based ways of meeting people, with Pickleball consistently emerging as one of the most requested options. Networking also came up repeatedly, particularly among professionals and expats in Saigon looking to meet people beyond their existing circles.
The Hangout page was updated so Sports, Networking and Road Trips were given the highest visibility in the category filters, followed by Tech and the remaining categories. Rather than treating every category equally, the order reflected the interests that came up most often during testing and made the most relevant starting points easier to reach.
One profile photo wasn't enough
The original profile required a single photo to establish identity and build trust. While the rest of the profile already provided rich information, founding members repeatedly raised that one photo wasn't enough to get a fuller sense of someone before connecting.
The profile page was therefore updated to require one primary photo while supporting up to four additional optional photos. This provided a richer visual introduction without turning profiles into another social media gallery.
Measurable Outcomes
Within 4.5 months, Nook progressed from an idea to a launch-ready native mobile application on iOS and Android. The product secured its first strategic partnership at prototype stage, with integration into an established digital ecosystem agreed as part of the partnership.
Post-launch performance metrics will be added as the first data becomes available.
Ongoing Product Development
Codebase ownership & long-term development
With the MVP now launch-ready, the focus shifts from rapid validation to long-term product evolution. The codebase has been migrated from Newly to GitHub, establishing a single source of truth. Development now continues directly within the codebase using an agentic coding tool in the terminal, enabling the product to evolve beyond the capabilities of an AI app builder while maintaining full control of the codebase.