Available for software engineering roles - backend systems, production reliability, cloud deployment, and AI-assisted workflows.

Connect
Responsible AI / Firebase MVPmvp

Pilly / MediMate Voice

Firebase-backed responsible-AI medication support MVP helping seniors respond to event-based reminders while caregivers see missed-dose, refusal, snooze, and help-request visibility.

Case file

Main outcome

Demonstrated event-based reminder flows for seniors and caregiver visibility.

Secondary signal

Separated AI response classification from medical advice or dosage decisions.

Evidence count

3 proof points

01

Problem

Seniors and caregivers need lightweight post-discharge medication support that can handle simple reminder responses without turning an AI prototype into a medical decision-maker.

02

Solution

Built a 12-hour Google AI hackathon MVP with Firestore-backed medication logs, Cloud Function-first workflows, Gemini response classification, deterministic fallback handling, and a Trusted Family Voice Reminder boundary.

03

Role

MVP product engineer · Responsible-AI workflow designer · Firebase backend implementer

Impact

  • Demonstrated event-based reminder flows for seniors and caregiver visibility.
  • Separated AI response classification from medical advice or dosage decisions.
  • Created a Firebase MVP shape that can be tested and extended without a custom backend first.

Technical Highlights

  • Firestore medication logs and response records
  • Cloud Function-first workflow orchestration
  • Gemini response classification with deterministic fallback
  • Missed-dose, refusal, snooze, and help-request event visibility
  • Trusted Family Voice Reminder boundary
  • Post-discharge medication support workflow

Proof

  • 12-hour Google AI hackathon MVP context
  • Public repo metadata and README notes
  • Explicit safety boundary requirements

Constraints

  • No diagnosis, dosage advice, or medical decision-making.
  • Urgent or ambiguous responses must route to static safety guidance or caregiver visibility.
  • AI classification is assistive and must have deterministic fallback behavior.

Limitations

  • Prototype only and not a medical device.
  • Requires clinical, caregiver, privacy, and accessibility review before real-world use.
  • Does not replace pharmacists, clinicians, emergency services, or prescribed care plans.

Roadmap

  • Add SAFETY_BOUNDARIES.md to the repo if missing.
  • Add tests for fallback classification, urgent phrases, refusal, help requests, and missed-dose alerts.
  • Add a short demo video showing the caregiver visibility flow.
All Case StudiesProof-of-Work Lab

Next conversation

Let's make the next system less fragile.

Open to software engineering roles across full-stack systems, platform and reliability work, workflow automation, and applied AI. I value teams where I can keep learning while contributing to real systems and clear delivery outcomes.

Also open to freelance or contract work across full-stack builds, practical automation, technical SEO, and cloud delivery.

githubgithublinkedinlinkedintwittertwitteremailemail
Get in Touch