Large Gmail inboxes are slow to triage one message at a time, but the data is deeply personal. The product needed to speed up cleanup without taking control away from the user or sending analysis off-device.
02On-device AI product
On-device AI that helps people reach Inbox Zero faster.
I designed and built a privacy-first Gmail cleaner that groups messages by sender, explains them with on-device AI, and enables fast, safe, reversible batch archiving.
- Live iOS product
- On-device AI
- Local analysis
- Reversible workflow
As founder and principal product engineer, I owned the product direction, experience design, architecture, engineering, validation, release controls, and launch.
MailBoom connects through secure Google OAuth, groups messages by sender, and runs AI categorization and summaries locally on-device. Users review Keep or Archive recommendations before applying reversible label changes; search and background processing support large inboxes.
The result is a live App Store product that turns a messy, high-volume workflow into a faster, privacy-conscious, user-controlled experience.