Roadmap & Algorithms — Boston Neuromind

📦 지금까지 만든 것 (인벤토리)What We've Built So Far (Inventory)What We've Built So Far (Inventory)What We've Built So Far (Inventory)Stuff We've Made So Far

🩺

Symptom Catcher (31개 AI 모듈, BPS-90, 양방향)Symptom Catcher (31 AI modules, BPS-90, bilingual)Symptom Catcher (31 AI modules, all bilingual)

v10.3.2 · LIVE

🧠

ADHD Catcher v11.0 — Adaptive Treatment EngineADHD Catcher v11.0 — Adaptive Treatment EngineADHD Catcher v11.0 (the smart one)

4,554 lines · LIVE

🧬

5개 백엔드 엔진 (S/V/τ/Q/π + Stage + Policy + Critical + Synthetic)5 backend engines (S/V/τ/Q/π + Stage + Policy + Critical + Synthetic)5 backend engines (the brains)

2,425 lines · LIVE

📚

매뉴얼 v11.0 (Clinician 15섹션 + Patient 7섹션, HTML+PDF)Manuals v11.0 (Clinician 15 sections + Patient 7 sections, HTML+PDF)Manuals v11.0 (long but comprehensive)

42 pages PDF · LIVE

🤖

BNM Companions 설계 페이지 (전체 비즈니스 + 6 모듈 + 4 페르소나)BNM Companions design page (full business + 6 modules + 4 personas)BNM Companions plan page (the big chatbot vision)

55 KB · LIVE

🛡️

Critical Factor Scanner (3-Tier + 8 위험 조합)Critical Factor Scanner (3-Tier + 8 dangerous combinations)Critical Factor Scanner (3-Tier + 8 dangerous combinations)Critical Factor Scanner (3-Tier + 8 dangerous combos)Safety scanner that catches dangerous stuff

525 lines · LIVE

🏷️

BNM Policy Library (8개 정책: 4 primary + 4 overlay)BNM Policy Library (8 policies: 4 primary + 4 overlay)BNM Policy Library (8 ready-to-use policies)

v11.0 · LIVE

🚀 개발 단계Development PhasesDevelopment PhasesDevelopment PhasesWhere We're At

완료된 단계 ✅, 현재 작업 🔥, 다음 단계 ⬜ Completed phases ✅, current work 🔥, upcoming ⬜ Completed phases ✅, current work 🔥, upcoming ⬜ Done ✅, in-flight 🔥, coming up ⬜ Done ✅, working on 🔥, coming up ⬜

Phase 0

2025-2026 Q12025-2026 Q1

✓

🩺 Foundation — Symptom CatcherFoundation — Symptom CatcherFoundation — The Big Screening Tool

31개 AI 모듈로 종합 정신건강 평가 시스템 구축. BPS-90 통합, 양방향 언어, 환자/임상가 모드. Built a comprehensive mental health assessment system with 31 AI modules. Integrated BPS-90, bilingual, patient/clinician modes. Built a comprehensive mental health assessment system with 31 AI modules. Integrated BPS-90, bilingual, patient/clinician modes. Built a full mental health assessment system with 31 AI modules. BPS-90 integrated, bilingual, patient/clinician modes. Built the big screening tool with 31 AI modules. Got BPS-90, both languages, patient and clinician modes.

31 AI modules (8 domains)
5-Lens analysis
Bilingual EN/KR
Critical Factor screening

✓ Complete

Phase 1

2026 Q22026 Q2

✓

🧠 ADHD Catcher v11.0 — Adaptive Treatment EngineADHD Catcher v11.0 — Adaptive Treatment EngineADHD Catcher v11.0 — The Smart One

5개 핵심 수학 함수 (S/V/τ/Q/π) 구축. AlphaGo 패러다임 적용 — 의사결정이 자동화되고 임상가가 검증. Built 5 core math functions (S/V/τ/Q/π). Applied AlphaGo paradigm — decisions auto-generated, clinician validates. Built 5 core mathematical functions (S/V/τ/Q/π). Applied the AlphaGo paradigm — decisions are auto-generated whilst the clinician validates. Built 5 core math functions (S/V/τ/Q/π). Used the AlphaGo idea — system makes the call, clinician signs off. Built 5 math functions. Like AlphaGo for therapy — system suggests, you approve.

5 core math functions (adaptive-core.js)
3-Stage system (Initial → Active → Integration)
Threshold-Free Architecture
Multi-Tag Policy System (8 BNM policies)
Synthetic Cohort Generator (literature-anchored)
Critical Factor Scanner (3-Tier + 8 combos)
Decision Cockpit UI

✓ Complete

Phase 2

2026 Q2-Q3 (현재)2026 Q2-Q3 (now)2026 Q2-Q3 (right now)

🔥

🤖💙 BNM Companions Phase 1 — ADHD Quick MVPBNM Companions Phase 1 — ADHD Quick MVPBNM Companions Phase 1 — ADHD MVP

Claude API 기반 ADHD 동반자 만들기. 4 페르소나, Pomodoro, Task breakdown, Mood log 모두 통합. 환자 1-3명 베타. Building Claude API-based ADHD companion. 4 personas, Pomodoro, task breakdown, mood log all integrated. Beta with 1-3 patients. Building a Claude API-based ADHD companion. 4 personas, Pomodoro, task breakdown, and mood log all integrated. Beta with 1-3 patients. Building the Claude-powered ADHD companion. 4 personas, Pomodoro, task breakdown, mood log — all in. Beta with 1-3 patients. Building the ADHD chatbot using Claude. 4 personas, Pomodoro, task breakdown, mood log. Testing with 1-3 patients.

Full design page (companions.html)
Claude API integration + ADHD prompts
4 persona system prompts
Pomodoro timer + cheering integration
Task breakdown chat
Mood + activity logging
Character selection UI (5 styles)
Background download system
Voice (browser TTS first)
Memory (localStorage → IndexedDB)

🔥 In Progress

Phase 3

2026 Q3-Q42026 Q3-Q4

3

🌟 BNM Specialty 3 모듈 완성 + Series A 준비BNM Specialty 3 Modules Complete + Series A Prep3 BNM Specialty Modules + Funding Prep

ADHD + Learning + Peak Performance 완성. 100명 환자 베타. ElevenLabs 음성 추가. Stripe 결제. Complete ADHD + Learning + Peak Performance. 100 patient beta. Add ElevenLabs voice. Stripe payments. Complete ADHD + Learning + Peak Performance. Beta with 100 patients. Add ElevenLabs voice. Stripe payments. Get ADHD + Learning + Peak Performance done. 100 patient beta. Slap on ElevenLabs voice. Stripe payments. Finish ADHD + Learning + Peak Performance modules. Test with 100 patients. Add fancy voice. Set up payments.

Learning Companion module
Peak Performance Companion module
100-patient beta
ElevenLabs voice integration
Stripe payment system
Series A pitch deck

⬜ Upcoming

Phase 4

2027 Q1-Q22027 Q1-Q2

4

🌍 Mass Market 확장 + Series A 실행Mass Market Expansion + Series A ExecutionMass Market Launch + Funding Round

Anxiety + Depression + PTSD 추가. Public app 출시. B2B 세라피스트 라이선스. Series A funding. Add Anxiety + Depression + PTSD. Public app launch. B2B therapist licensing. Series A funding. Add Anxiety + Depression + PTSD. Public app launch. B2B therapist licensing. Series A funding. Add Anxiety + Depression + PTSD. Public launch. B2B therapist licensing. Series A funding. Add Anxiety + Depression + PTSD. Public launch. License to therapists. Raise Series A.

Anxiety Companion module
Depression Companion module
PTSD Companion module
Public app launch
B2B Solo Therapist licensing
Series A funding execution

⬜ Upcoming

Phase 5

2027+2027+

5

🚀 글로벌 확장 + Anthropic 파트너십Global Expansion + Anthropic PartnershipGoing Global + Anthropic Deal

다국어 (KR/EN/JP/CN/SP). White-label B2B. Series B. 카테고리 정의자 지위. Multi-language (KR/EN/JP/CN/SP). White-label B2B. Series B. Category-defining position. Multi-language (KR/EN/JP/CN/SP). White-label B2B. Series B. Category-defining position. Multi-language (KR/EN/JP/CN/SP). White-label B2B. Series B. Defining the whole category. Many languages. White-label for clinics. Series B. Define the whole category.

⬜ Upcoming

🧬 5개 핵심 수학 함수5 Core Math Functions5 Core Mathematical Functions5 Core Math FunctionsThe 5 Math Functions

AlphaGo의 평가 패러다임을 임상 뉴로피드백 3D 공간 (State × Time × Context)에 적용 Applied AlphaGo's evaluation paradigm to clinical neurofeedback's 3D space (State × Time × Context) Applied AlphaGo's evaluation paradigm to clinical neurofeedback's 3D space (State × Time × Context) Took AlphaGo's evaluation paradigm and applied it to NF's 3D space (State × Time × Context) AlphaGo's evaluation idea, applied to NF's 3D space (State × Time × Context)

마스터 방정식: Master Equation: Decision = π(S, T, C)

S(p, t)

상태 함수State Function

환자 → 13차원 벡터Patient → 13D vectorPatient as 13D point

시점 t의 환자를 5 QEEG + 3 척도 + 2 기능 + 1 자가보고 + 1 관찰 + 1 시간 = 13차원 벡터로 변환. Converts patient at time t into 5 QEEG + 3 scales + 2 function + 1 self-report + 1 observation + 1 time = 13D vector. Converts the patient at time t into a 13D vector (5 QEEG + 3 scales + 2 function + 1 self-report + 1 observation + 1 time). Turns the patient at time t into a 13D vector — 5 QEEG + 3 scales + 2 function + 1 self-report + 1 observation + 1 time. Turns patient into a 13D point — 5 brain + 3 scales + 2 function + 1 feeling + 1 observation + 1 time.

V(S, C)

가치 함수Value Function

"이 상태가 얼마나 좋은가""How good is this state"

현재 상태가 목표에 얼마나 가까운지 0~1 점수. 단계별 가중치로 (초기엔 자가보고, 중기엔 QEEG, 후기엔 기능). Scores how close current state is to target, 0-1. Stage-aware weights (Initial: subjective, Active: QEEG, Integration: function). Scores how close the current state is to the target, 0-1. Stage-aware weights (Initial: subjective, Active: QEEG, Integration: function). Scores how close current state is to target, 0-1. Stage-aware weights — Initial uses self-report, Active uses QEEG, Integration uses function. Score from 0-1 — how close to the target. Different things matter at different stages.

V = 1 - Σ wᵢ(stage) · dist(Sᵢ, target)

τ(p, n)

궤적 함수Trajectory Function

"어떻게 변하고 있는가""How is patient changing"

최근 N 세션의 변화를 분석 — 방향, 속도, 가속도, 변동성, 고원 감지, 신뢰도. Analyzes change over recent N sessions — direction, velocity, acceleration, volatility, plateau detection, confidence. Analyses change over recent N sessions — direction, velocity, acceleration, volatility, plateau detection, confidence. Looks at change over recent N sessions — direction, velocity, acceleration, volatility, plateau detection, confidence. Looks at recent sessions — what direction, how fast, getting faster?, stable?, plateau?

Q(S, A, C)

행동 가치 함수Action Value Function

"이 행동의 예상 결과""Expected outcome of action"

합성 + 실제 코호트로부터 학습 — 비슷한 환자가 이 행동 했을 때 결과 통계. 신뢰도 + 샘플 크기 같이 반환. Learned from synthetic + real cohort — statistics of outcomes when similar patients took this action. Returns confidence + sample size. Learnt from synthetic + real cohort — statistics of outcomes when similar patients took this action. Returns confidence + sample size. Learnt from synthetic + real cohort — what happened when similar patients took this action. Returns confidence + sample size. Learns from similar patients — what happened when they tried this action. Includes confidence.

π(S, C) ⭐

정책 함수 — THE COREPolicy Function — THE CORE

"최종 결정 함수""The final decision function"

매 세션 호출되는 함수. Critical Scan → Q 평가 → Trajectory 통합 → 최적 행동 선택 → 설명 생성. 신뢰도 + 대안 + 예측 trajectory 모두 반환. Called every session. Critical scan → evaluate Q → integrate trajectory → select best action → generate reasoning. Returns confidence + alternatives + predicted trajectory. Called every session. Critical scan → evaluate Q → integrate trajectory → select best action → generate reasoning. Returns confidence + alternatives + predicted trajectory. Runs every session. Critical scan → evaluate Q → integrate trajectory → pick best action → explain. Returns confidence + alternatives + predicted trajectory. Runs every session. Safety check → score actions → look at trends → pick best → explain why. Includes alternatives + prediction.

π = argmax_A Q(S, A, C) + trajectory + safety

⚖️ 11가지 설계 원칙 (Phase 1 결정 사항)11 Design Principles (Phase 1 Decisions)11 Rules (How We Built It)

알고리즘 구축 중 11번의 본질적 결정으로 형성된 시스템 철학 System philosophy formed through 11 fundamental decisions during algorithm construction System philosophy formed through 11 fundamental decisions during algorithm construction System philosophy from 11 key decisions during the build 11 big decisions that shaped the system

1

진척도 5축5-Axis Progress

QEEG > 척도 > 목표 > 자가보고 > 관찰 (객관 → 주관) QEEG > Scales > Goals > Self-report > Observation (objective → subjective)

2

단계별 가중치Stage-Based Weights

초기엔 자가보고, 중기엔 QEEG, 후기엔 목표 우세 Initial: subjective, Active: QEEG, Integration: goals

3

하이브리드 단계 전환Hybrid Stage Transition

세션 가드레일 + 데이터 자동 + 임상가 거부권 Session guardrails + data-driven auto + clinician veto

4

Threshold-FreeThreshold-Free

모든 숫자는 정책 안 (코드에 하드코딩 X) All numbers in policies (no hardcoded thresholds) No magic numbers in code — all in editable policies

5

다차원 태그 정책Multi-Tag Policies

Goal/Diagnosis/Demographic/Style/Overlay 자유 조합 Free combination of Goal/Diagnosis/Demo/Style/Overlay

6

Critical Factors 별도Critical Factors Separate

자살/위험 약물 등은 일반 알고리즘 X, 전문 처리 Suicide/dangerous combos bypass general algorithm Dangerous stuff handled separately, not by general logic

7

합성 코호트 부트스트랩Synthetic Cohort Bootstrap

Literature anchor + 시뮬레이션 → 콜드스타트 해결 Literature anchors + simulation → cold-start solved

8

자가 보정Self-Correction

실제 환자 늘면 합성 가중치 자동 감소 Synthetic weight auto-decreases as real patients grow

9

4-Layer 검증4-Layer Validation

통계 (auto) + 자가보정 (auto) + 임상가 검토 + 문헌 Statistical (auto) + self-correction (auto) + clinician review + literature

10

XAI (설명 가능)XAI (Explainable)

모든 결정에 신뢰도 + 근거 + 대안 자동 생성 Every decision: confidence + reasoning + alternatives

11

하이브리드 충돌 해결Hybrid Conflict Resolution

기본 = 안전한 값, 임상가 override 가능 Default = safer value, clinician override allowed

⚙️ 기술 스택Tech StackWhat We Use

🤖

Claude API

대화/추론 AIConversation/reasoning AI

📊

Mitsar EEG

19채널 QEEG19-channel QEEG

🧠

NeuroGuide

Normative DBNormative DB

⚡

HBImed + HBI

ERP + functional normsERP + functional norms

🎯

BioGraph Infiniti

NFB deliveryNFB delivery

💻

JavaScript ES6+

전체 시스템Entire system

☁️

AWS S3 + CloudFront

호스팅 + CDNHosting + CDN

🔄

GitHub Actions

자동 배포Auto-deploy

🎯 Symptom Catcher v2.0 설계 결정 (2026-04-23)Symptom Catcher v2.0 Design Decisions (2026-04-23)Symptom Catcher v2.0 Decisions (2026-04-23)

인간의 결정 오류 문제 해결 + AI의 진정한 활용으로 재설계 Redesigned to solve human decision-making errors + truly leverage AI capabilities Rebuilt to fix human decision-making errors and truly use AI properly Rebuilt to fix human decision errors and really use the AI

🚨 해결한 문제The Problem We Solved

❌

인간이 모듈 선택Human chooses modules

환자가 "이 31개 중 해당되는 것 고르세요" = 자기 진단 지식 필요 + 편향 많음 Patient picks from 31 modules = requires self-diagnosis knowledge + biased Patient picks from 31 modules = needs to self-diagnose + lots of bias

❌

AI 능력 미활용AI underutilized

설문지 관리 기능만 — Claude의 pattern recognition, DSM reasoning 미사용 Just survey management — Claude's pattern recognition, DSM reasoning unused

❌

500+ 문항 부담500+ questions burden

31 모듈 × 5-15문항 = 압도적, 환자 지침 31 modules × 5-15 questions = overwhelming, patient fatigue 31 modules × lots of questions = way too much, patients give up

✅ v2.0 아키텍처 (A+B+C 조합)v2.0 Architecture (A+B+C Combo)

🅰️

Triage Role (법적 보호)Triage Role (Legal Protection)

"AI는 진단 안 함 — 경로 안내만" — 비의사 practitioner scope 준수, NCMHCE 기준 "AI doesn't diagnose — just routes" — non-physician practitioner scope, NCMHCE aligned "AI just routes, doesn't diagnose" — keeps us within non-physician scope

+

🅱️

Initial Screening (합리적 기초)Initial Screening (Reasonable Foundation)

핵심 scale + AI 대화 = 의심 진단 목록 + 우선순위 Core scales + AI conversation = list of suspected diagnoses + priorities

+

🅲

Differential Dx (AI 강점)Differential Dx (AI's Strength)

DSM-5-TR + Claude reasoning = 진단 확률 + 공존진단 + 감별진단 자동 DSM-5-TR + Claude reasoning = diagnosis probabilities + comorbidity + differential auto

→

📐 v2.0 5단계 흐름v2.0 5-Step Flow

Step 1. 자연어 대화Natural Conversation

"무엇 때문에 오셨어요?" — Claude가 증상 자연스럽게 추출 "What brings you here?" — Claude extracts symptoms naturally
Step 2. 똑똑한 모듈 선택Smart Module Selection

Claude가 자동으로 적절한 모듈 제안 (환자 승인 또는 조정) Claude auto-suggests relevant modules (patient confirms or adjusts)
Step 3. 최소 핵심 평가Minimal Key Assessments

각 모듈에서 가장 discriminative한 3-5문항만 (기존의 ~20%) Only 3-5 most discriminative questions per module (~20% of original)
Step 4. AI 차별 진단AI Differential Diagnosis

Claude + 문헌 + DSM-5-TR → 확률 목록 + 공존진단 + Critical Factor 안전검사 Claude + literature + DSM-5-TR → probability list + comorbidity + Critical Factor safety check
Step 5. 경로 추천Routing Recommendation

"ADHD Catcher로 가시겠어요?" — 경로 안내 + 임상가 최종 확인 "Shall we go to ADHD Catcher?" — routing + clinician final confirmation

🤖 Conversational Companion은 ADHD Catcher로 이동Conversational Companion Moves to ADHD Catcher

핵심 통찰: "동반자 = 치료 내내 계속" → Symptom Catcher는 1-2회 평가, ADHD Catcher는 20-30회 여정. 동반자는 ADHD Catcher 안에 있어야 함! Key insight: "Companion = continuous through treatment" → Symptom Catcher is 1-2 assessments, ADHD Catcher is 20-30 session journey. Companion must live inside ADHD Catcher! Key insight: "Companion = stays with you through treatment." So it belongs in ADHD Catcher (20-30 sessions), not Symptom Catcher (1-2 assessments).

📚 컨텐츠 커버리지 전략 (3-Tier)Content Coverage Strategy (3-Tier)

T1

Curated CoreCurated Core

여보야 임상 경험 + 핵심 문헌 + DSM-5-TR = 500-1000 검증된 패턴 Clinical experience + key literature + DSM-5-TR = 500-1000 validated patterns

T2

AI-Generated 확장AI-Generated Augmentation

Claude가 variation 생성 → 임상가 검증 = 10,000+ 케이스 패턴 Claude generates variations → clinician validates = 10,000+ patterns

T3

Runtime ClaudeRuntime Claude

edge cases는 실시간 Claude 추론 + Critical Scanner 안전망 Edge cases via real-time Claude + Critical Scanner safety net

💡 핵심 원칙Core Principle

"100% 사전 커버 불가능 → Core + AI runtime + 안전망 = 실용적 완성도" "100% pre-coverage impossible → Core + AI runtime + safety net = practical excellence" "Can't cover 100% upfront → Core + AI + safety net = good enough to work"

✅ v2.2 완료 (2026-04-22)v2.2 COMPLETED (2026-04-22)v2.2 Done (2026-04-22)

Symptom Catcher v2.2: 3-Role + 4-Methodology + 3-Tier Synthesis + Windows/Mac 호환 Symptom Catcher v2.2: 3-Role + 4-Methodology + 3-Tier Synthesis + Windows/Mac compatible

✓

3-Role 선택3-Role Selection

환자 직접 / 세라피스트 동반 / 전문가 수동 — 명확한 역할 Patient / Therapist / Expert — clear role separation

✓

4-Methodology4-Methodology

Hybrid(기본) / Evidence / AI / Print+Email Hybrid(default) / Evidence / AI / Print+Email

✓

3-Tier Synthesis3-Tier Synthesis

Keyword + Claude + Critical Scanner 통합 Keyword + Claude + Critical Scanner integrated

✓

Print + EmailPrint + Email

질문지 PDF 출력 + mailto 이메일 배포 Questionnaire PDF print + mailto email distribution

✓

Admin ConfigAdmin Config

BNM 전용 — methodology + 가중치 설정 BNM-only — methodology + weight settings

✓

Win/Mac 호환Win/Mac Compat

UTF-8 BOM, Safari 인쇄, 클립보드 폴백 UTF-8 BOM, Safari print, clipboard fallback

✓

Evidence 보존Evidence Preserved

31 모듈 + BPS-90 + normative DB 그대로 31 modules + BPS-90 + normative DB intact

✓

매뉴얼 업데이트Manual Updated

Clinician + Patient, KO + EN, HTML + PDF Clinician + Patient, KO + EN, HTML + PDF

🔥 다음 우선순위 (이번주~다음주)Next Priorities (This week ~ Next week)What's Next

1

🤖 Claude API + ADHD Companion MVP

단일 페이지 데모. Claude API 직접 호출, ADHD 특화 프롬프트, 4 페르소나 시스템. 작동하는 첫 데모. Single-page demo. Direct Claude API call, ADHD-specific prompts, 4 persona system. First working demo. Single-page demo. Direct Claude API call, ADHD-specific prompts, 4 persona system. First working demo. Single-page demo. Direct Claude API call, ADHD-specific prompts, 4 persona system. First working demo. Single-page demo. Direct Claude API, ADHD prompts, 4 personas. First working demo.
2

⏰ Pomodoro + Task Breakdown 통합Pomodoro + Task Breakdown Integration

집중 세션 + 챗봇 응원, 일일 task 분해 대화, 진행 시각화. Focus sessions + chatbot cheering, daily task breakdown chats, progress visualization. Focus sessions with chatbot cheering, task breakdown chats, progress vis.
3

📊 Mood + Activity 로깅Mood + Activity Logging

매일 자동 로깅, 패턴 자동 발견, 임상가 알림. Daily auto-logging, pattern auto-discovery, clinician alerts.
4

🎨 캐릭터 선택 + 배경 시스템Character Selection + Background System

5가지 캐릭터 스타일 (Anime/Cartoon/Realistic/Mascot/Abstract), 무료 5 + 유료 배경 다운로드. 5 character styles (Anime/Cartoon/Realistic/Mascot/Abstract), 5 free + paid background downloads.
5

🎵 음성 (시작은 브라우저 TTS)Voice (start with browser TTS)

브라우저 TTS로 시작 → 나중에 ElevenLabs 통합. Start with browser TTS → integrate ElevenLabs later. Start with browser TTS → upgrade to ElevenLabs later.
6

💾 메모리 시스템 (localStorage → IndexedDB)Memory System (localStorage → IndexedDB)

대화 기록 영구 저장, 패턴 추출, 환자별 personal AI 진화. Persistent conversation history, pattern extraction, per-patient personal AI evolution.

💎 최종 비전Ultimate VisionThe Big Vision

Boston의 작은 클리닉에서 시작해, 진짜 케어할 수 있는 AI의 새 카테고리를 만들고, 전 세계 수억 명의 정신건강을 변화시킨다. Starting from a small clinic in Boston, creating a new category of AI that truly cares, transforming the mental health of hundreds of millions worldwide. Starting from a small clinic in Boston, creating a new category of AI that truly cares, transforming the mental health of hundreds of millions worldwide. Starting from a small clinic in Boston, building a new category of AI that actually cares, transforming mental health for hundreds of millions worldwide. Started from a small Boston clinic. Building a new kind of AI that actually cares. Going to change mental health for millions of people.

— Boston Neuromind LLC

Boston Neuromind 로드맵 & 알고리즘 Boston Neuromind Roadmap & Algorithms Boston Neuromind Roadmap & Algorithms Boston Neuromind Roadmap & Algorithms BNM Where We're Going & How It Works

📦 지금까지 만든 것 (인벤토리)What We've Built So Far (Inventory)What We've Built So Far (Inventory)What We've Built So Far (Inventory)Stuff We've Made So Far

🚀 개발 단계Development PhasesDevelopment PhasesDevelopment PhasesWhere We're At

🩺 Foundation — Symptom CatcherFoundation — Symptom CatcherFoundation — The Big Screening Tool

🧠 ADHD Catcher v11.0 — Adaptive Treatment EngineADHD Catcher v11.0 — Adaptive Treatment EngineADHD Catcher v11.0 — The Smart One

🤖💙 BNM Companions Phase 1 — ADHD Quick MVPBNM Companions Phase 1 — ADHD Quick MVPBNM Companions Phase 1 — ADHD MVP

🌟 BNM Specialty 3 모듈 완성 + Series A 준비BNM Specialty 3 Modules Complete + Series A Prep3 BNM Specialty Modules + Funding Prep

🌍 Mass Market 확장 + Series A 실행Mass Market Expansion + Series A ExecutionMass Market Launch + Funding Round

🚀 글로벌 확장 + Anthropic 파트너십Global Expansion + Anthropic PartnershipGoing Global + Anthropic Deal

🧬 5개 핵심 수학 함수5 Core Math Functions5 Core Mathematical Functions5 Core Math FunctionsThe 5 Math Functions

⚖️ 11가지 설계 원칙 (Phase 1 결정 사항)11 Design Principles (Phase 1 Decisions)11 Rules (How We Built It)

⚙️ 기술 스택Tech StackWhat We Use

🎯 Symptom Catcher v2.0 설계 결정 (2026-04-23)Symptom Catcher v2.0 Design Decisions (2026-04-23)Symptom Catcher v2.0 Decisions (2026-04-23)

🚨 해결한 문제The Problem We Solved

✅ v2.0 아키텍처 (A+B+C 조합)v2.0 Architecture (A+B+C Combo)

Triage Role (법적 보호)Triage Role (Legal Protection)

Initial Screening (합리적 기초)Initial Screening (Reasonable Foundation)

Differential Dx (AI 강점)Differential Dx (AI's Strength)

📐 v2.0 5단계 흐름v2.0 5-Step Flow

🤖 Conversational Companion은 ADHD Catcher로 이동Conversational Companion Moves to ADHD Catcher

📚 컨텐츠 커버리지 전략 (3-Tier)Content Coverage Strategy (3-Tier)

✅ v2.2 완료 (2026-04-22)v2.2 COMPLETED (2026-04-22)v2.2 Done (2026-04-22)

🔥 다음 우선순위 (이번주~다음주)Next Priorities (This week ~ Next week)What's Next

💎 최종 비전Ultimate VisionThe Big Vision