Problem
Inboxes overwhelm.
Modern inboxes flood users with tasks, notifications, updates, unread anxiety, and fragmented workflows.
Available for AI PM roles · 2026
Shivani Vaidya.
Building AI-native products that reduce chaos.
I'm a product-minded software engineer who enjoys building AI-native products that reduce chaos instead of adding to it. Over the past 10+ years, I've worked on cloud, AI, and distributed systems across cybersecurity and data platforms, but somewhere along the way, a conversation about Ikigai made me realize the parts of my work I loved most were understanding people, shaping experiences, and helping teams make sense of complexity.
Most recently, I've been building MailBuddy, an AI inbox assistant inspired by a simple question: why do modern inboxes feel emotionally exhausting? I've also worked on AI-powered coaching experiences, storytelling-driven engagement, and personalization systems designed to make technology feel a little more thoughtful.
Outside of work, I'm a singer, storyteller, and mom to an energetic toddler — which probably explains both my appreciation for calm user experiences and my strong feelings about notification overload.
Currently obsessed with AI workflows, product UX, and why unread counts feel emotionally aggressive.
Featured project
MailBuddy — From Inbox to Action
AI-powered inbox assistant focused on reducing cognitive overload from email.
Core insight
Email is no longer email.
Inboxes have become workflow systems — blending communication, tasks, recommendations, updates, and operational coordination into one stream.
Solution
From stream to signal.
MailBuddy transforms overwhelming inboxes into prioritized actions, compressed updates, workflow summaries, and actionable intelligence.
What it does.
AI task extraction
Pull actionable to-dos out of long threads, automatically.
Semantic grouping
Cluster emails by intent, not just by sender or label.
Thread summaries
Compress 40-message chains into a 3-line situational read.
Workflow prioritization
Surface what blocks others and what blocks you.
Voice assistant
Triage and respond hands-free, even between meetings.
Smart updates dashboard
Newsletters, releases, status — compressed into one view.
Product artifacts
Behind the product.
The thinking, research, and decisions that shaped MailBuddy — documented as a real PM artifact.
Featured artifact
Product Specification
How MailBuddy evolved from an observation about inbox overload into an AI-native semantic intelligence layer for Gmail.
- Problem Definition
- User Research Insights
- Product Vision
- MVP Scope
- Product Decisions
- Success Metrics
Explore in Figma
lilac-coyote-21147049.figma.site
Product Specification
Open live document
Interactive product spec hosted on Figma Sites.
Product thinking
Notes from building with users in the loop.
01
Unread counts create psychological stress.
Numbers train the brain to chase zero, not meaning. The cost is attention, not minutes.
02
Not all emails deserve equal attention.
A receipt and a board update share a row. The interface flattens what shouldn't be flat.
03
Email clients organize messages, not intent.
Folders and labels describe storage. People want to know what to do next.
04
AI should reduce cognitive load, not add noise.
If summaries become another stream to read, the product has failed its premise.
Selected work · Case study
ZenAI.Coach
Product Manager · AI-powered coaching platform
Worked as the sole Product Manager on a small cross-functional team building AI-powered coaching experiences focused on storytelling, personalization, emotional engagement, and guided self-reflection.
Storytelling & engagement
Turning storytelling into product strategy
Recognized that users connected more deeply with relatable stories than abstract concepts. Proposed and introduced story-driven coaching journeys that eventually became one of the most positively received parts of the product experience.
User research & personas
Designing for emotional context
Defined user and AI coach personas to create personalized coaching journeys tailored to different life situations, motivations, and emotional needs. Used user interviews and behavioral feedback to shape engagement strategies and personalization experiences.
The Quiet SeekerThe Overloaded HeartThe Analytical FeelerThe Unanchored Soul
AI personalization
Digital Twin personalization system
Designed the product vision for a Digital Twin system that helps AI coaches remember context, personalize guidance, and adapt coaching journeys over time.
Memory systems
Semantic retrieval
Contextual personalization
Adaptive journeys
Product × engineering
Bridging product thinking and AI systems
Contributed to architecture and AI workflow discussions around personalization systems, memory design, semantic retrieval, and feature prioritization by bridging product and engineering perspectives.
AI workflows→ Systems thinking→ Prioritization
Voice-first coaching
Voice-First AI Coaching
Helped design a coaching experience that blended conversational AI, storytelling, journaling, and guided meditation into a seamless daily practice.
Conversational AIVoice ExperiencesCoaching FlowsGuided Reflection
Product discovery & user research
Listening before building
Conducted user interviews and feedback sessions to uncover engagement patterns, retention challenges, usability pain points, and opportunities to improve the coaching experience. Used these insights to shape product direction, storytelling strategy, personalization efforts, and future roadmap decisions.
User InterviewsProduct DiscoveryCustomer FeedbackExperimentation
Reflection
What I learned building AI coaching experiences
Building emotionally engaging AI experiences requires much more than good models. It requires narrative, pacing, trust, emotional resonance, and thoughtful product design.
Beyond work
A full life makes for better products.
Beyond the screen
Indian Classical Vocalist
Rooted in storytelling, performance, and emotional connection — years of training in raag and taal that quietly shape how I build.
Life outside the IDE
Mom to an energetic toddler
Successfully potty-trained a very opinionated toddler in 10 days. Behavior change, incentives, and iterative experimentation matter a lot there too.