Crafting AI solutions and food delivery experiences
At bbdeeplearning.systems, I stepped into a dynamic environment where innovation met rapid execution. As part of a small but ambitious team, we thrived on creativity, collaboration, and the challenge of building transformative solutions under tight timelines. This was not the setting for static roles—I wore many hats, balancing design, development, and research to deliver impactful products.
FeynChat, one of our flagship projects, sought to revolutionize how users interact with AI by integrating multiple AI models into one seamless platform. It was ambitious and technically challenging but reflected our belief in creating tools that adapt to user needs rather than the other way around.
The ONDC food delivery solution, part of India’s Open Network for Digital Commerce, aimed to simplify food ordering through a blend of conversational AI and traditional interfaces, making the experience seamless and accessible for everyone. The dual focus of leveraging cutting-edge AI technologies and addressing practical, real-world problems defined the spirit of our work at bbdeeplearning.systems. Over these two years, I had the privilege of contributing to these projects, shaping products that blended innovation with usability.
2 years 3 months (Jun 2022 - Aug 2024)
2 years 3 months (Jun 2022 - Aug 2024)
2 years 3 months (Jun 2022 - Aug 2024)
2 years 3 months (Jun 2022 - Aug 2024)
UI/UX Designer, UX Researcher, Front End Developer
UI/UX Designer, UX Researcher, Front End Developer
UI/UX Designer, UX Researcher, Front End Developer
UI/UX Designer, UX Researcher, Front End Developer
Figma, Github, Jira, VS Code, React
Figma, Github, Jira, VS Code, React
Figma, Github, Jira, VS Code, React
Figma, Github, Jira, VS Code, React
FeynChat unified tools like ChatGPT and Claude into a multi-model platform tailored to diverse user needs. While the interface drew inspiration from familiar layouts, the design process focused on solving unique challenges, like seamless model switching, refining prompts with SmartPrompt, and converting PDFs into audio with DocTalk.
My design work began with early prototypes exploring various layouts, where continuous iteration shaped intuitive and modular workflows. Collaborating closely with engineers, I refined React components in real-time, addressing technical challenges while ensuring the platform delivered a seamless and user-friendly experience. This hands-on approach allowed me to bridge design and development effectively.
ONDC App
The ONDC food delivery solution is a work in progress that combines conversational AI with traditional interfaces to simplify food ordering and shopping experiences. Our vision is to create an accessible platform that bridges the gap between advanced AI and real-world services, redefining how users engage with technology in their daily lives.
To inform the design, I conducted research comparing interaction patterns between conversational agents and search engines for grocery shopping. The study highlighted key insights: conversational interfaces offered a more natural experience for users with limited tech expertise, while search engines provided faster results for those familiar with traditional workflows.
My research revealed a fascinating divide: inexperienced users gravitated toward conversational agents because they brought everything together in one place, making grocery ordering feel simple and accessible. On the other hand, experienced users leaned on the control and familiarity of search engines but couldn’t ignore the speed and ease that conversational tools offered. This pointed us toward a hybrid solution—combining the best of both worlds.
sprint 2: user experience
sprint 3: api and prototyping
sprint 4: user testing results
Homepage (First Screen):
Users liked the centralized dashboard but found it cluttered.
Product Selection (Second Screen):
Product details like ratings were helpful, but comparing options in the chat interface was tricky.
Empty Cart (Third Screen):
The empty cart message was visually clear but lacked actionable recommendations.
Order Confirmation (Last Two Screens):
Users appreciated the conversational tone but wanted a clearer cost breakdown.
Integrating personalized recommendations to support user decision-making.
Optimizing the checkout flow to ensure transparency and reduce friction.
Enhancing product comparison through clearer layouts and expandable details.
FeynChat unified tools like ChatGPT and Claude into a multi-model platform tailored to diverse user needs. While the interface drew inspiration from familiar layouts, the design process focused on solving unique challenges, like seamless model switching, refining prompts with SmartPrompt, and converting PDFs into audio with DocTalk.
My design work began with early prototypes exploring various layouts, where continuous iteration shaped intuitive and modular workflows. Collaborating closely with engineers, I refined React components in real-time, addressing technical challenges while ensuring the platform delivered a seamless and user-friendly experience. This hands-on approach allowed me to bridge design and development effectively.
FeynChat unified tools like ChatGPT and Claude into a multi-model platform tailored to diverse user needs. While the interface drew inspiration from familiar layouts, the design process focused on solving unique challenges, like seamless model switching, refining prompts with SmartPrompt, and converting PDFs into audio with DocTalk.
My design work began with early prototypes exploring various layouts, where continuous iteration shaped intuitive and modular workflows. Collaborating closely with engineers, I refined React components in real-time, addressing technical challenges while ensuring the platform delivered a seamless and user-friendly experience. This hands-on approach allowed me to bridge design and development effectively.
FeynChat unified tools like ChatGPT and Claude into a multi-model platform tailored to diverse user needs. While the interface drew inspiration from familiar layouts, the design process focused on solving unique challenges, like seamless model switching, refining prompts with SmartPrompt, and converting PDFs into audio with DocTalk.
My design work began with early prototypes exploring various layouts, where continuous iteration shaped intuitive and modular workflows. Collaborating closely with engineers, I refined React components in real-time, addressing technical challenges while ensuring the platform delivered a seamless and user-friendly experience. This hands-on approach allowed me to bridge design and development effectively.
let's build something impactful.
let's build something impactful.
let's build something impactful.
let's build something impactful.
let's build something impactful.
