Case study · AI / voice ML · Parenting tech

Cappella: Android cry detection for an MIT-founded baby-care startup

Contract senior Android engineer (Jun–Oct 2024) for Cappella — “Empowering Modern Parenting with Sustainable Tech.” Finished the Android product surfaces, wired voice ML on the baby-device path, and helped land the Google Play launch. Credited by name on the founder’s Android launch post.

Cappella LinkedIn article cover: Hear me out — AI sound analysis for infant cries

Cappella baby cry interpreter — Hungry Cappella Select Unit — parent and baby profiles Cappella live baby monitor streaming

Product

Cappella (Cambridge, MA; founder Apolline Deroche, MIT Sloan) was an AI parenting brand: a two-phone baby monitor that reuses devices parents already own, plus an AI cry translator and routine tracking. The company built a hospital-collected dataset of 5,500+ labelled infant cries across six need classes — hungry, tired, in pain, lower-body discomfort, upper-body discomfort, and respiratory disease signals — with clinicians in the labelling loop.

Positioning on LinkedIn: sustainable tech for modern parenting, CES 2024 demo lineage, and an all-in-one app so parents are not juggling separate sleep / feeding / diaper tools.

Problem

iOS led; Android needed an experienced engineer to close the remaining task list and ship. A substantial Kotlin codebase existed, but launch depended on baby-device cry detection, care-tracking UX, audio library work, Java ML library integration, and a long tail of auth / profile / rotation bugs.

What I shipped

Public product features (launch messaging)

Outcomes

References

Product UI imagery from the official Cappella App Store listing (archived Feb 2024); LinkedIn posts cited above for launch narrative and team credit.

Hire for AI / ML Android work → Contract engagements