Malvika Jain - Portfolio

Portfolio

Malvika Jain

About

I build technology at the intersection of engineering and design: systems that are both technically sophisticated and have thoughtful interfaces. My work has ranged from embedded sensor networks at MIT Media Lab to visualization tools for Qualcomm's machine learning team to founding Shmood, where I've raised over $600,000 building multimodal AI systems for creative agencies. As a 23-year-old solo founder, I've learned that the hardest problems aren't purely technical. They're about creating interfaces people trust and want in their daily lives. This started early. I grew up building smart furniture: tables that lit up when someone sat down, paintings that responded to touch. I was never certain where art ended and engineering began, and I still don't draw that line. Whether I'm coding backend systems, designing interfaces, or teaching physics in rural Kenya, I'm interested in one thing: making complex technology feel approachable and useful.

Projects

MIT Media Lab Responsive Environments Group (AR + sensor network for plant health)

Augwe (AR fashion as creative interfaces)

Gisty (engineering education subscription boxes)

Scenery (engineer + designer at video editing software startup)

Shmood (AI for design feedback)

Art & Furniture (smart furniture + interactive art)

MIT Media Lab Responsive Environments Group

ESP32 Arduino Firebase Unity C# AR Design PCB Design Sensor Networks

During my freshman summer at MIT Media Lab's Responsive Environments Group, I developed a wireless sensor network for monitoring soil health. I designed compact sensor nodes using ESP32 microcontrollers to capture real-time light, moisture, and temperature data, which was transmitted to a Firebase database. To make this environmental data more accessible, I built an augmented reality visualization in Unity that transformed quantitative metrics into intuitive, interactive visual representations, bridging the gap between raw sensor data and human understanding.

AR Gardening Aid demonstration video

AR avatar overlaid on laptop showing plant type and real-time status

Problem

Gardening is hard. Beginners and young kids lack the intuition to understand what plants need. Light, moisture, and temperature interact in complex ways, and existing tools simply display numbers without context or actionable guidance. This disconnect prevents successful plant growth and limits access to home agriculture.

Insights

The key is translating numbers into something people understand at a glance. Visual metaphors work better than raw data. Mapping sensor readings to avatar behaviors creates immediate understanding. By overlaying digital information directly onto physical plants through AR, I reduced the cognitive load needed to assess plant health. The goal: make invisible environmental factors visible.

Solution

I built a complete system with four components: ESP32 sensor nodes that monitor soil conditions, a Firebase database that aggregates real-time data, a Unity AR app for visualization, and algorithms that translate sensor values into intuitive avatar behaviors. I designed compact, power-efficient systems to keep sensors small and unobtrusive. The system works in household planters and scales to larger agricultural applications.

Complete system flow: ESP32 transmits sensor data via WiFi to Firebase, which feeds Unity AR app that maps data to avatar actions and colors

Hardware design: ESP32 microcontroller with sensors on breadboard (left) and data mapping diagram showing how sensor values translate to avatar behaviors (right)

Three avatar states showing different environmental conditions: (1) Little water/light/low temp = yellow/white joints/watering (2) Too much light/low temp = red/white joints/picking (3) Too much light/high temp/low water = red/red joints/watering

Potential applications: system designed for household planters with scalability to larger indoor farming installations

Takeaway: This taught me that better sensing and interfaces help humans understand data, but I kept wondering about systems that could act on insights independently.

Augwe

Product Design Mobile Development iOS AR QR Integration Go-to-Market Fashion Tech

I founded Augwe (using the reseach at MIT Media Lab) to create beautiful designs and experiences that could invoke important conversations. Starting as a voter registration campaign, it evolved into something bigger: AR-linked clothing that reimagines garments as interfaces for amplifying creative voices and social causes. By merging physical and digital worlds, we transformed how people express themselves through what they wear. I built the complete product ecosystem: merchandise, website, iPhone app, and social media presence. The response exceeded expectations: Augwe became a hit on Duke's campus, and Bumble paid us to create scannable merchandise promoting their events. This experience taught me that technical capability means nothing without genuine user adoption. People have to want to invite technology into their lives.

Augwe demonstration video

Augwe brand identity and merchandise on Duke's campus

Problem

Online conversations are static snapshots that don't capture how we actually feel. Fashion and social media existed as separate domains, neither facilitating meaningful in-person dialogue or authentic expression.

Insights

What we wear is already a conversation starter. I saw potential in combining fashion and technology to create a new platform for digital communication that capitalizes on the physical interactions we have with clothing every day.

Solution

We built a complete platform: designed clothing and bags, developed an iPhone app using Unity and Vuforia for image-triggered AR experiences, created a website, and managed social media. When people scanned garments, they unlocked stories and connections. The platform gained traction quickly. Bumble paid us to create scannable hoodies promoting campus events, proving the concept's commercial viability.

Sample customer interaction showing the complete user experience

Brand inspiration and mood board

Augwe merchandise designs featuring AR-enabled clothing and accessories

Augwe website: homepage and team page showcasing the platform and people behind the project

iPhone app interface with AR scanning and social media presence showing community engagement

Bumble collaboration overview

Bumble collaboration: initial design concepts and mockups for scannable event merchandise

Design iterations and variations exploring different visual approaches for campus event promotion

Final designs and production-ready merchandise for Bumble campus events

Takeaway: This taught me that technical capability means nothing if people won't invite the technology into their lives.

Gisty

Product Design Education Subscription Model Community Building Engineering

For my senior thesis, I researched why teenage girls and artists turn away from engineering, not because they lack ability, but because of how the field is presented. I created Gisty, a design-first subscription service that packages engineering like trendy makeup and clothing brands. Each month, users receive beautifully designed boxes with tools, materials, and QR codes linking to short film-like tutorials. Later, I brought this approach to 60 girls at WISER in Kenya, teaching physics.

Gisty subscription box introduction video

Gisty subscription box designed to make engineering feel approachable and exciting

Problem

Lack of relatable content in engineering tutorials. There's a stigma around what an engineer should look like. Although engineering is fundamentally creative, this aspect isn't highlighted, and many who would develop an interest are left out.

Insights

There's value in simplicity when designing accessible tutorials. Packaging information like trendy consumer brands makes it appealing to people not in tech. Market research and testing on social media showed that shorter videos with trending formats work better.

Solution

Created a subscription service shipping tools and instructions for monthly projects. Each box contains components (Arduino, LEDs, breadboards, wires) and QR codes linking to short film-like tutorials. Boxes are carefully designed in Adobe Illustrator, inspired by clothing and artwork.

Solution overview: YouTube tutorials and subscription service model

Target market: teenagers and their parents

Box designs and printmaking process - each box carefully designed in Adobe Illustrator

Prototype subscription box with engineering components and tutorials

Takeaway: This taught me that how we communicate technology matters as much as what we build.

Scenery

Real-time Video In-browser Editing Design Systems Figma JavaScript React AWS Engineering + Design Cross-functional

My first job after Duke was at Scenery, a startup building real-time in-browser video editing, later acquired by Adobe. As the youngest member of a 20-person team, I worked as both a designer and engineer. The team quickly began relying on me to bridge gaps between technical and creative sides. I built design systems in Figma and led two major product initiatives: redesigning the keyboard shortcuts menu and expanding our titles library.

Project 1: Keyboard Shortcuts Redesign

Problem

We offered numerous keyboard shortcuts but it was difficult for users to find the specific shortcut they were looking for. The existing menu lacked discoverability and efficient browsing capabilities.

Solution

Redesigned the keyboard shortcut menu with search functionality and improved browsing structure. Created an intuitive interface that allows users to quickly find and learn shortcuts, improving editing efficiency.

Success Criteria

Measure increased discoverability through user searches, reduced time to find shortcuts, and improved adoption of keyboard shortcuts in editing workflows.

Keyboard shortcuts menu redesign showing improved organization and search capabilities

Final implementation with categorized shortcuts and search functionality

Project 2: Titles Library Expansion

Problem

Our title library was too limited with not enough options to choose from. We wanted to offer bespoke titles with customization flexibility (colors, fonts, positioning) while maintaining quality and ease of use.

Solution

Added new bespoke titles to the library and created a custom text title option. Developed a scalable process for creating high-quality titles with full customization capabilities including text, colors, fonts, and positioning.

Success Criteria

Track usage of custom text vs bespoke titles, measure retention of users utilizing titles, analyze which specific titles are most popular, and gather qualitative feedback on title selection quality.

Project planning and success criteria for titles library expansion

Titles library interface with browsing and customization options for text, fonts, and positioning

Takeaway: This experience showed me that my ability to translate between disciplines was rare and valuable, and that there was an opportunity to use AI to solve communication problems in creative work at scale.

Shmood

Founder AI Agents RAG Architecture Multimodal AI JavaScript React Next.js AWS Backend Product Strategy Fundraising Team Building

As a solo founder, I raised over $600,000 and hired a team of engineers and designers from Apple and Palantir, including Cornell alumni and my previous manager. Drawing on my experience building intuitive interfaces for complex systems, we focused on a problem I had experienced repeatedly: creative agencies struggled to gather clear, structured feedback from clients about creative work. We built a platform that used AI agents to help clients articulate design critiques. Our system analyzed conversations and designs, prompted clarifying questions, and automatically synthesized feedback into reports using Retrieval-Augmented Generation (RAG). The technical challenge was managing large amounts of data and maintaining low latency while keeping costs manageable.

Shmood platform overview and demo

Problem

Design agencies struggled to gather clear feedback from clients about their artistic vision. Clients couldn't articulate their preferences precisely, leading to endless revision cycles. Designers needed better tools to understand what clients actually wanted, not just what they said they wanted. The communication gap cost significant time and money.

Insights

Clients are saturated with software tools, every new platform risks becoming another thing to manage. I learned that AI could bridge the gap between what clients see and what they can articulate, but only if it maintains low latency, minimizes hallucinations, and stays authentic to the creative's actual voice. Technical reliability determines whether creatives will actually adopt it.

Solution

We built AI agents that analyzed designs and client conversations in real-time, asking clarifying questions that helped clients express preferences more precisely. The system automatically synthesized feedback into comprehensive case studies. We also created RAG-based personalized chatbots that maintained authentic voice while providing recommendations. The technical challenge was balancing latency, cost, and hallucination minimization.

Shmood platform interface showing project dashboard and portfolio management for collecting client feedback

Using AI to generate case studies

AI-generated summaries from design feedback

Presenting work and gathering feedback

Mobile interface flow showing complete user journey from project creation to feedback collection

Responsive design system adapting seamlessly across desktop, tablet, and mobile devices

Takeaway: We could analyze text and images, but the contextual richness that shapes human taste was invisible to software. That realization pushed me toward robotics.

Art & Furniture

Woodworking Painting Furniture Design Arduino Circuits Laser Cutting Drawing Craftsmanship Sewing

I grew up building smart furniture: tables that lit up when someone sat down, paintings that responded to touch. Since early childhood, my parents taught me that my environment was my responsibility. If the trash was full, I took it out. If the living room needed a coffee table, I built one. These projects blend functionality with aesthetic expression, creating objects that bring warmth and intentionality to everyday spaces. I was never certain where art ended and engineering began, and I've come to realize maybe there's no boundary at all.