Early workProgress snapshot

Zahimar App

Face-aware assistant concept — “who is this, again?”

An exploratory Flutter build around face recognition ideas (TensorFlow / ML Kit territory) to help people feel less awkward in social recall—not a shipped consumer product, but a sincere lab for camera, models, and trust.

Zahimar App preview

The idea (one sentence)

Use on-device intelligence as a gentle prompt—not a surveillance story—so remembering names feels supportive instead of invasive.

Where this sits in my arc

Three beats — curiosity first, ethics second, product craft third.

Spark · then

ML as a magic trick phase

I chased the “wow” of TensorFlow / ML Kit working inside Flutter: camera frames in, labels out. The scope was intentionally narrow because I was still learning how brittle vision features can be in real rooms and real lighting.

Lesson · what stuck

Clever is cheap; consent is not

Face tech forces you to ask who owns the embedding, what fails closed, and what copy you show users. That mindset stuck—today I treat sensitive flows with the same seriousness whether the sensor is a camera or a health wearable.

Trajectory · toward

From demo to dependable UX

I still like ambitious interfaces—but now they sit on top of clear architecture, BLoC-style state, and integration boundaries. Zahimar was the sketch; later products are where the ergonomics and scale caught up.

Stack (as listed)

  • Flutter
  • TensorFlow / ML Kit exploration
  • Camera + inference UX

For hiring teams

This entry is proof I do not hide awkward early experiments. If your culture rewards learning in public and people who tighten scope as they mature, you are looking at the same engineer who later ships structured, multi-surface apps—just further along the same vector.