Yukti Arora

Product Designer & Former Designer Founder

Yukti Arora

Product Designer & Former Designer Founder

Work

About

Press & Writing

Resume

Work

About

Press & Writing

Resume

I specialize in zero to one enterprise products.

My experience makes me fluent in driving clarity in ambiguous problem spaces, with a strong focus on balancing business and user needs.

I’m experimenting novel interaction patterns for AI, designing experiences that feel less like tools and more like collaborators.

Notable things about me.

Notable things
about me.

Nurtured people and built innovative, collaborative cultures.
Shared ideas on play, dance & AI, and innovation - from cozy workshops to global stages.

Nurtured people and built innovative, collaborative cultures.
Shared ideas on play, dance & AI, and innovation - from cozy workshops to global stages.

Nurtured people and built innovative, collaborative cultures.
Shared ideas on play, dance & AI, and innovation - from cozy workshops to global stages.

Work. Examples of my Thinking + Making

Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template (Only first time)
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template (Only first time)
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template (Only first time)
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template (Only first time)
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.

Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.

Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template (Only first time)
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template (Only first time)
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template (Only first time)
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.
Step 1: Splitting sheets
I worked closely with the engineering team to figure out the feasibility of automatically splitting the PDF, during upload and within the browser.
Step 1.2: Data Extraction Through Template (Only first time)
The next step was naming the sheets and adding additional metadata to it. From my research, I knew all this data was present in the "Title Block".
Step 2: Review Drawings
This step ensures accuracy by enabling human review of OCR results before upload - maintaining accuracy and control.

Buildsys Drawing Upload

I used AI Automation to turn a 2 day construction task into a 10 minute workflow.

Founder & Design Lead

2019

Problem

Uploading construction drawings and metadata was a slow, error-prone process—taking up to 2 days per project. Misfiled or outdated versions led to costly rework and delays, creating a major pain point across the construction lifecycle.

Solution

We built Automated Drawing Upload—a 10-minute workflow that automated sheet splitting, naming, metadata extraction, sorting, and version control. What took 2 days for 100+ drawings could now be done in minutes.

I led the effort, balancing tight engineering bandwidth, no ML capability, and complex stakeholder needs. By aligning the experience with users’ existing mental models, we shipped a solution that worked out of the box and felt instantly familiar.

Impact

Faster

100x

Improved efficiency by reducing task time from 2 days to 20 minutes.

Staff Costs

20%

Achieved a 20% reduction in staff costs.

Less Rework

30%

Reduced rework costs by 30%.

TL;DR

Case study

Buildsys Meeting Minutes

I used data analysis and usability testing to transform the process of taking meeting minutes, reducing churn rate by 42%.

Founder & Design Lead

2020

Problem

The Meeting Minutes feature—critical to Buildsys adoption—was facing high early churn. 45% of users dropped off within 14 days, and only 27% completed a meeting record. The process was click-heavy, inflexible, and made it hard to import past notes.

Solution

I led the end-to-end redesign of Meeting Minutes, using a data-informed approach to identify drop-off points and guide iteration.

I introduced a WYSIWYG editor with keyboard shortcuts and slash commands to reduce friction and better match users’ mental models.

I collaborated closely with engineering to balance scope with technical constraints, and with sales to ensure the updated experience addressed pain points surfaced during demos and customer calls.

We launched the solution through a phased rollout, refining it with feedback from pilot accounts before a full release.

Impact

Fewer Clicks

Reduced task creation from 6+ clicks to 1 click.

Task Completion

52%

Minutes publish rate increased from 27% to 41%.

Retention

42%

Meetings module retention over 30 days went up by 42%.

TL;DR

Case study

Problem

Solution

I led the end-to-end redesign of Meeting Minutes, using a data-informed approach to identify drop-off points and guide iteration.

I introduced a WYSIWYG editor with keyboard shortcuts and slash commands to reduce friction and better match users’ mental models.

We launched the solution through a phased rollout, refining it with feedback from pilot accounts before a full release.

Impact

Fewer Clicks

Reduced task creation from 6+ clicks to 1 click.

Task Completion

52%

Minutes publish rate increased from 27% to 41%.

Retention

42%

Meetings module retention over 30 days went up by 42%.

TL;DR

Case study

Natya.AI

I crafted the vision for an AI-powered practice companion for dancers that provides real time voice and visual feedback.

Product Designer

2024

Problem

80% of dance is practice—and 80% of practice is correction.

Dancers record their sessions to visualize movement and self-correct. But reviewing is cumbersome, time-consuming, and breaks the flow of practice.

The two core challenges:

Lack of live feedback
Review fatigue—scrubbing recordings for mistakes takes more time than the practice itself

Solution

Natya.AI is a practice companion for Bharatanatyam dancers that uses multimodal AI to provide real-time voice and visual feedback on live recordings.

It reduces review fatigue and helps dancers focus on meaningful, actionable correction.

Impact

Review Time

50%

7/7 users reported a 50% reduction in video review time, enabling more focus on practice.

Practice Efficiency

95% of dancers found Natya “very helpful” for identifying and fixing mistakes.

Watch Talk

Delivered a talk on the interplay between Dance & AI

The voice prompts felt like having my Guru in the room. It will completely change the way I approach solo practice.

- Varshini, Bharatnatyam Dancer & Performer

This is the best application of AI I have seen so far.

- Sam Potts, Designer @ Apple

TL;DR

Problem

80% of dance is practice—and 80% of practice is correction.

Dancers record their sessions to visualize movement and self-correct. But reviewing is cumbersome, time-consuming, and breaks the flow of practice.

The two core challenges:

Lack of live feedback
Review fatigue—scrubbing recordings for mistakes takes more time than the practice itself

Solution

Natya.AI is a practice companion for Bharatanatyam dancers that uses multimodal AI to provide real-time voice and visual feedback on live recordings.

It reduces review fatigue and helps dancers focus on meaningful, actionable correction.

Impact

Review Time

50%

7/7 users reported a 50% reduction in video review time, enabling more focus on practice.

Practice Efficiency

95% of dancers found Natya “very helpful” for identifying and fixing mistakes.

Watch Talk

Delivered a talk on the interplay between Dance & AI

The voice prompts felt like having my Guru in the room. It will completely change the way I approach solo practice.

- Varshini, Bharatnatyam Dancer & Performer

This is the best application of AI I have seen so far.

- Sam Potts, Designer @ Apple

TL;DR

Instant, non-intimidating, and personalized AI-based support
Call, chat, and breathing exercises use familiar interaction patterns for a warm, personalized experience.

Elpis

I simplified a complex subject matter through iterative UX testing and intuition.

Design Lead

2026

Problem

Miscarriage is an isolating experience, and mental health support for women post-loss is often inaccessible, impersonal, or absent.

Existing tools fail to offer timely, emotionally attuned, and private support when it’s most needed.

Solution

I designed Elpis, an AI-supported mental health companion that offers self-paced, non-intimidating support for women after miscarriage.

I led sensitive user research, prototyping, and product direction—starting with three concepts, and refining the final experience using Google’s HEART metrics to balance feasibility, desirability, and emotional resonance.

The final solution included CBT-based chat, therapist discovery, and peer support—designed to feel safe, warm, and personal.

Impact

Yukti! I just wanted to say that your ability to lean in, ask questions, and hold space on such a sensitive topic was amazing and it showed in the project you developed that felt so tailored to our experience. You're brilliant!

- Research Participant

TL;DR

Problem

Miscarriage is an isolating experience, and mental health support for women post-loss is often inaccessible, impersonal, or absent.

Existing tools fail to offer timely, emotionally attuned, and private support when it’s most needed.

Solution

I designed Elpis, an AI-supported mental health companion that offers self-paced, non-intimidating support for women after miscarriage.

The final solution included CBT-based chat, therapist discovery, and peer support—designed to feel safe, warm, and personal.

Impact

- Research Participant

TL;DR

In case you’re curious.

Here’s my story.

I studied Industrial Design at Ohio State University, where I discovered that my greatest passions and abilities were at the intersection of visual thinking, digital product design, and business strategy. These are skills I learned by doing at my startups. I realized I wanted to work on problems that excite me and could impact the world.

In search of impact, I founded Buildsys, a construction productivity SaaS. At Buildsys, I learned what it takes to manage a creative culture, built teams that approach problems from a human-centered perspective, crafted business, sales, and marketing strategies, and designed a product that is easy to use and adopt.

From 2017 to 2022, I scaled Buildsys from an idea to a successful enterprise SAAS business with clients in 10 cities across India and exited the company in early 2022.

I created Wazo Space Station, an interactive digital dollhouse set in outer space- designed for play, storytelling, and fun for kids ages 3-7.

I'm in love with the power of play and curious about the interplay of design, technology, and play as a way to have an impact on how we learn, tell stories, spark curiosity, build connections, and collaborate.

All of these experiences makes me fluent in driving clarity in ambiguous problem spaces. I also realized it’s about tradeoffs - that balances business and user needs.

In 2024, I graduated with an MFA in HCI from the School of Visual Arts, New York, where my thesis explored the interplay between dance and AI.

When not doing a serious play, I am probably making a pretty looking salad, drawing with friends, climbing, dancing, snowboarding or dreaming of the outdoors.