HALO

LLLT Hair-Growth Cap, Hardware PM Spec

“A PRD for something you wear on your head, with lasers.”

outcomes

Full PRD: problem → goals → non-goals → user stories → requirements → metrics
Hardware build spec: optical system, safety interlocks, BLE firmware state machine
Phased roadmap with 5 phases and explicit exit gates
Engineering handoff ready
Dose target: ~3–4 J/cm² at scalp, 10 min/day

pm skills

PRD writinggoals vs non-goalsuser story mappingsafety as product requirementphased roadmap with exit gatesengineering handoffhonest scope definition

Problem

Commercial LLLT caps for hair loss cost $700+. They are closed black boxes, no visibility into dose, pulse profile, or adherence. For protocol-driven users already running a disciplined regimen, the opacity is the actual pain point. Not the price. The control.

What I Built

A complete hardware product specification: PRD, hardware build spec, and a phased roadmap with explicit exit gates, ready for engineering handoff or hardware partner.

Approach

Started with honest framing. LLLT is adjunctive, not foundational. The PRD explicitly states what this product is not, not a medical device, not a replacement for the real levers, not a high-power device. Non-goals written before goals.

Safety requirements written as hard outcomes. "Emitters physically cannot fire off-head or in a fault state" is a requirement, not a guideline.

State	Entry Condition	Exit Condition
IDLE	Power on	User initiates session
SELF_TEST	Session start	All checks pass
READY	Self-test pass	Scalp contact confirmed
RUNNING	Contact confirmed	Timer complete or fault
COMPLETE	Timer done	User removes device
FAULT	Any safety breach	Manual reset required

Roadmap uses exit gates, not timelines. Phase 0 exit = one diode lights and holds steady output. Phase 1 exit = every fault condition provably kills output with emitters stubbed. You don't move on until the gate clears.

PM Skills Applied

This is the purest PM artifact in the portfolio. Requirements written as outcomes. Safety as a first-class requirement, not an afterthought. Goals and non-goals scoped tightly so an engineer can make decisions without escalating. Full handoff package: PRD + build spec + roadmap, all cross-referenced.

Technical Depth

Firmware state machine: IDLE → SELF_TEST → READY → RUNNING → COMPLETE/FAULT. APC via monitor photodiode feedback loop. ESP32-S3 MCU + BLE. Constant-current driver banks (PWM). Scalp-contact safety interlock (capacitive/IR-reflectance). 650–660 nm emitter matrix (~80–120 diodes). Fluence target: ~3–4 J/cm² at scalp, 10 min/session.

Lessons

Lesson	Why It Matters
Non-goals are as important as goals	They prevent scope from becoming a medical device startup
Safety requirements must be outcomes, not intentions	"Should not" is not a spec
Exit gates over timelines	In hardware, "it works" is the only honest milestone

The Story

(Hardware not yet built, full spec and roadmap complete)

HALO exists to answer a question I kept asking about every commercial LLLT device: what's the actual dose? Nobody publishes it. The PRD starts there, if you can't measure it, you can't trust it, and if you can't trust it, it's not a protocol tool, it's a wellness product. Everything else flows from that. The spec took a week. The build will take months. That's what a PRD is for.