§ WHAT Research Trio — What does this Skill Group do?
- Skill group
01 · B— Three optional research skills used before Brainstorm or PRD- Market Research
- · Phase 01 · Agent: Mary (Analyst)
- Domain Research
- · Phase 01 · Agent: Mary (Analyst)
- Technical Research
- · Phase 03 · Agent: Dev / Tech Lead
- Output pattern
_bmad-output/planning-artifacts/research/{type}-{topic}-research-{date}.md- Hard prerequisite
- Web search must be enabled — no fallback to training data alone
The Research Trio covers three orthogonal failure modes. Market Research validates who the customers are and what the competitive landscape looks like. Domain Research maps the industry's vocabulary, regulations, and expert mental models. Technical Research evaluates technology feasibility, stack alternatives, and architecture patterns. Each skill is independent — run one, two, or all three depending on what the team doesn't yet know.
three-tracks Three Research Tracks
Track Overview
Research Trio · 01·B
| Track | Skill | What it answers | Agent | Phase |
|---|---|---|---|---|
| Market | bmad-market-research |
Who are the customers? What do competitors offer? What is the market size and growth rate? What are the unmet needs? | Mary (Analyst) | 01 · Analysis |
| Domain | bmad-domain-research |
What vocabulary do experts use? What regulations apply? What is the value chain? What are the industry's mental models? | Mary (Analyst) | 01 · Analysis |
| Technical | bmad-technical-research |
Is the architecture feasible? What are the stack alternatives? What are the integration patterns? What are the performance and security constraints? | Dev / Tech Lead | 03 · Dev |
decision-matrix When to Run Which Skill
Decision Matrix
Research Trio · skill selection
| Situation | Skill to run |
|---|---|
| Team unfamiliar with this industry | Domain Research — learn the vocabulary, regulations, and mental models before writing a word of the PRD |
| New customer segment never served before | Market Research — validate who the customers are, what they pay, and who already serves them |
| Novel technology or architecture to evaluate | Technical Research — confirm feasibility, compare stack options, surface integration patterns |
| All three above simultaneously | Run Market → Domain → Technical in order (see Execution Order below) |
| Known domain, existing segment, decided stack | Skip Research Trio — proceed directly to Brainstorming or PRD |
order Recommended Execution Order
When running more than one skill, the order matters — each skill's output informs the next.
1
Market Research · first
Understand the commercial landscape
Run first to confirm the segment exists, understand who the customers are, and see what competitors already offer. The findings define scope for Domain Research — you know which regulations and vocabulary are relevant because you know which market you're entering.
2
Domain Research · second
Map the industry's expert mental models
Run second to learn the regulations, vocabulary, and value chain of the domain the market research identified. The regulatory findings directly affect which technology constraints Technical Research needs to evaluate.
3
Technical Research · third
Confirm feasibility against real constraints
Run last — after you know what market you're building for and what regulations apply. Technical Research then evaluates architecture options against those real constraints rather than in the abstract.
§ WHY Why Research Before PRD?
Writing a PRD before researching is writing fiction. Each research track removes a specific class of blind spots that are expensive to discover in Phase 4 or after launch.
blind-spots Three Blind Spots — One Per Track
| Dimension skipped | Blind spot | Typical failure |
|---|---|---|
| Skip Market Research | Wrong customer assumptions, no competitor visibility, pricing in the dark | Pricing users won't accept; discovering a competitor had the same feature two years ago; building for a persona that does not exist in the actual customer base |
| Skip Domain Research | Wrong vocabulary, missing regulations, expert mental models unknown | Healthcare app using marketer language that doctors reject; HIPAA violation discovered at launch; product design that conflicts with a fundamental industry workflow |
| Skip Technical Research | Wrong feasibility estimate, vendor lock-in, missed stack alternatives | Phase 4 discovering the chosen database doesn't meet compliance requirements; epic rewrite because the architecture doesn't scale; integration that took 2 weeks estimated takes 6 months |
Compounding effect: All three blind spots compound. A team that doesn't know the market can't write a useful domain glossary. A team that doesn't know the domain can't write accurate regulatory constraints for the architect. An architect without constraints designs for the wrong system. Research Trio breaks this compounding cycle before it starts.
§ HOW Before Calling the Skills
Everything to have ready before invoking any Research Trio skill. Gathering this upfront produces richer, more targeted output — the agent can only surface constraints it knows about.
Universal prerequisite: Web search must be active in your Claude Code session. All three research skills require live web search — they do not fall back to training data alone. The agent will confirm search availability in Step 1 of each skill.
bmad-market-research Market Research — What to Prepare
Preparation Checklist
bmad-market-research · before session
| Item | What's needed | Priority |
|---|---|---|
| Research topic | Precise statement: which customer segment, which geography, which use-case. The more specific, the richer the output. | Required |
| Research goals | 2–3 concrete questions to answer: "What do I want to know?" e.g. "What is the TAM and who are the top 5 competitors?" | Required |
| Scope | Broad (whole market) or deep (one segment)? Global or one region? | Required |
| Known competitors | Names already known — Step 5 will go deep on these plus discover unknown players. | Helpful |
| Hypotheses to test | "I think our primary customer is X" — research confirms or refutes. | Optional, high-value |
Skills That Help Market Research Go Smoothly
- JTBD thinkingUnderstanding why a customer "hires" a product — not demographics but jobs. Helps you ask sharper research questions and interpret the output more precisely.
- Pain vs. symptom"Customers complain about slow loading" = symptom. "Can't close deals without live data" = pain point. Knowing the difference helps you frame the research question correctly.
- Competitive intelFamiliarity with App Store / G2 reviews, LinkedIn job postings, annual reports. These are the sources the agent draws from most — knowing how to read them speeds interpretation.
Market Research — 6-Step Process
1
Initialization
Confirm topic, goals, and scope
No web search yet. Agent confirms the 3 setup items and presents them back for approval before any research begins.
2
Customer Behavior & Segments · 4 parallel searches
Behavior patterns, demographics, psychographics, adoption drivers
Four simultaneous web searches covering different angles of the customer. Sources: industry association reports, government data, product review platforms, journalism.
3
Pain Points & Needs · 4 parallel searches
Challenges, frustrations, unmet needs, adoption barriers
Four simultaneous searches on what the customer struggles with. Primary source: G2 / Capterra / App Store reviews, Reddit threads, community discussions.
4
Decision Journey · 4 parallel searches
Decision processes, buying criteria, journey mapping
Four simultaneous searches on how customers evaluate and choose solutions. Sources: analyst reports, vendor case studies, B2B buying research.
5
Competitive Analysis
Key players, market share, positioning, SWOT
Deep dive on competitors — both known players and newly discovered ones. Sources: SEC filings, earnings calls, product websites, G2 comparisons.
6
Research Completion
Strategic synthesis → 11-section document
Synthesizes all prior steps into a structured document with executive summary, customer profiles, competitive map, and strategic recommendations. User selects [C] to save.
bmad-domain-research Domain Research — What to Prepare
Preparation Checklist
bmad-domain-research · before session
| Item | What's needed | Priority |
|---|---|---|
| Domain statement | "We are building [product] in the [X] industry" — specific industry name, not "tech" or "enterprise". | Required |
| Known unknowns | What do you know you don't know about this domain? List them explicitly — the agent will target these. | Required |
| Geographic scope | Regulations vary by region. US, EU, SEA, global, or a specific country. | Required |
| Regulatory hunches | "I think this involves HIPAA" or "I don't know what applies" — both are valid starting points. | Helpful |
| Time budget | Complex regulated domains (healthcare, finance, insurance) take 2–4 hours for a thorough session. | Planning |
Skills That Help Domain Research Go Smoothly
- Industry classificationKnowing whether the domain is regulated vs. high-growth tech vs. emerging helps calibrate depth expectations before starting.
- Value chain awarenessUnderstanding that regulations often sit at specific handoff points in the value chain speeds up Step 4 (Regulatory Focus).
- Regulatory literacyKnowing the difference between a regulation (legal requirement) and a standard (voluntary best practice) prevents misclassifying findings in the output.
Domain Research — 6-Step Process
1
Domain Scope Confirmation
Confirm 5 research areas — no web search yet
Agent defines the 5 areas: Industry Analysis, Competitive Landscape, Regulatory Focus, Technical Trends, Synthesis. User approves before research begins.
2
Industry Analysis · 4 parallel searches
Market size, growth dynamics, industry structure, trends
Four simultaneous searches. Sources: industry association reports, government statistical agencies, analyst press releases, business journalism.
3
Competitive Landscape · 4 parallel searches
Key players, strategies, business models, ecosystem
Four simultaneous searches on who is in the industry and how they compete. Sources: company websites, LinkedIn, Crunchbase, earnings calls.
4
Regulatory Focus · 3 searches
Regulations, standards, data privacy requirements
Three searches targeting regulations, industry standards, and data privacy rules. Sources: government websites, regulatory bodies, legal commentary.
5
Technical Trends · 3 searches
Emerging tech, digital transformation, future outlook
Three searches on technology adoption in the industry. Sources: analyst reports, industry association tech publications, conference proceedings.
6
Research Synthesis
10-section document with strategic recommendations
Synthesizes all steps into a structured domain reference covering industry overview, regulations, competitive landscape, and strategic insights. User selects [C] to save.
bmad-technical-research Technical Research — What to Prepare
Preparation Checklist
bmad-technical-research · before session
| Item | What's needed | Priority |
|---|---|---|
| Technology question | Specific: "Compare GraphQL vs REST for our use case" — not "research about APIs." The more precise, the more actionable the output. | Required |
| Architecture context | The Architecture Document from Phase 3, or a description of the current/planned architecture. | Required |
| Constraints | Language/framework constraints, team expertise limits, licensing requirements, compliance constraints from Domain Research. | Required |
| Comparison candidates | List specific technologies to compare — the agent will parallel-analyze each one explicitly. | Required |
| Decision deadline | Technical Research feeds into Solution Design — knowing the deadline helps the agent calibrate depth vs. speed. | Helpful |
Skills That Help Technical Research Go Smoothly
- Architecture literacyKnowing the difference between monolith, microservices, and serverless patterns means the agent can skip introductory framing and go straight to your specific trade-off question.
- Integration thinkingUnderstanding that most systems integrate with existing services helps define the constraint space before the session starts.
- Performance baselineHaving rough performance requirements (e.g. "p99 < 200ms", "10k concurrent users") lets the agent evaluate scalability patterns against real targets.
Technical Research — 6-Step Process
1
Technical Scope Confirmation
Confirm 5 areas: Architecture, Implementation, Stack, Integration, Performance
No web search yet. Agent defines the research scope across 5 technical dimensions. User approves before research begins.
2
Technology Stack Analysis · 4 parallel searches
Languages, frameworks, databases, cloud services
Four simultaneous searches comparing technologies on the candidate list. Sources: official docs, Stack Overflow surveys, GitHub activity, Hacker News discussions.
3
Integration Patterns · 4 parallel searches
API patterns, protocols, microservices, event-driven
Four simultaneous searches on integration and communication patterns relevant to the architecture context.
4
Architectural Patterns · 3 searches
Architecture patterns, design principles, scalability
Three searches targeting known architectural patterns and their trade-offs at the scale and context described in the architecture document.
5
Implementation Research · 3 searches
Adoption strategies, DevOps practices, operational considerations
Three searches on how teams implement and operate the technologies in practice. Sources: engineering blogs, conference talks, case studies.
6
Research Synthesis
12-section document with implementation roadmap and risk assessment
Synthesizes all prior steps into a structured technical reference covering stack analysis, architecture patterns, integration, security, and strategic recommendations. User selects [C] to save.
toolkit Applied Toolkit reference
Detailed coverage profiles for each of the three research skills. Click a card to jump to the relevant section.
Market Research · bmad-market-research
Data Sources & Sector Coverage
Where the agent sources market data, what it can and cannot access, and which industry sectors produce the richest output.
→ View Coverage
Domain Research · bmad-domain-research
Domain Taxonomy — 25 Domains
All 25 industry domains the agent covers, with tier classification and trigger keywords that activate each domain's specialized research depth.
→ View Taxonomy
Technical Research · bmad-technical-research
Technology Domain Coverage
All technology domains covered across three depth tiers: Expert (15 domains), Strong (9 domains), Adequate (6 domains).
→ View Coverage
bmad-market-research Market Data Sources
The agent has no private database. Every data point comes from live web search at session time.
Accessible Sources
Accessible — Live Web Search
bmad-market-research · data sourcing
| Source type | Examples | What you get |
|---|---|---|
| Market research free tiers & summaries | Statista (free charts), Grand View Research / MarketsandMarkets press releases, Gartner/Forrester public summaries | Headline market-size numbers, CAGR figures, top-level segment breakdowns |
| Industry association reports | NRF (retail), HIMSS (healthcare IT), CB Insights fintech, Shopify/BigCommerce state-of-commerce, CBRE (real estate) | Sector-specific statistics published freely; often the most reliable first-party data |
| Public company filings | SEC 10-K / 10-Q, earnings call transcripts, annual reports | Revenue by segment, customer counts, competitor commentary in their own words |
| Business journalism | Reuters, Bloomberg (free), Financial Times, Forbes, TechCrunch, VentureBeat | Funding rounds, M&A events, competitive moves, expert commentary |
| Product review platforms | G2, Capterra, Trustpilot, App Store / Play Store reviews | Customer pain points, feature comparisons, NPS proxies, complaints at scale |
| Government statistical agencies | US Census, BLS, EU Eurostat, World Bank Open Data, OECD | Demographics, employment by industry, trade volumes, GDP by sector |
| Academic abstracts | Google Scholar, PubMed, SSRN, arXiv | Research findings — abstract is free even when the full paper is paywalled |
| Community & social platforms | Reddit (r/[industry]), LinkedIn job postings, Hacker News | Authentic user frustrations, tech adoption signals, hiring priorities as market signals |
NOT Accessible
Not Accessible — Paywalled / Proprietary
bmad-market-research · limitations
| Source type | Examples |
|---|---|
| Full paid analyst reports | Gartner Magic Quadrant full document, Forrester Wave full report, IDC Market Tracker subscriptions |
| Proprietary BI platforms | Bloomberg Terminal, PitchBook full data, Nielsen Syndicated Data |
| Private company financials | Revenue, margin, customer counts from non-public companies without public disclosures |
| Real-time transactional data | Live pricing, live sales volumes, live inventory, real-time competitor dashboards |
| Paywalled academic journals | Full-text papers behind institutional paywall (abstract only) |
Calibration note: Numbers cited by the agent (e.g., "$12.4B by 2028, CAGR 14.3%") come from a publicly available summary — not a purchased analyst report. Treat all market size figures as directional estimates suitable for PRD context-setting, not for investor decks or strategic commitments.
Sector Coverage — Where Output Is Richest
Sector Coverage Tiers
bmad-market-research · output quality by sector
| Tier | Sector | Why data is rich | What you get |
|---|---|---|---|
| Best | Technology / SaaS | Gartner & Forrester public summaries, G2/Capterra, Crunchbase, Statista tech charts | Market size with CAGR, competitor feature comparisons, pain points from reviews, funding landscape |
| Best | Consumer Mobile / Apps | App Store charts, Sensor Tower public summaries, data.ai reports, social analytics, Reddit | Category rankings, DAU/MAU benchmarks, retention benchmarks, monetization comparisons |
| Best | E-commerce / Retail | Shopify State of Commerce, Adobe Digital Economy Index, NRF reports, eMarketer summaries | GMV figures, conversion benchmarks, category growth rates, consumer behavior shifts |
| Best | Fintech / Payments | Central bank publications, regulatory filings, CB Insights, Visa/Mastercard presentations | Payment volumes, adoption rates, regulatory requirements, market breakdown by segment |
| Strong | Healthcare / MedTech | CDC/WHO/NIH data, FDA approval databases, hospital association reports, PubMed abstracts | Patient statistics, regulatory pathway descriptions, market sizes, tech adoption in care settings |
| Strong | Media / Streaming | Publicly reported subscriber counts (earnings), Nielsen press releases, MPA/RIAA data | Subscriber counts, time-spent metrics, content investment, advertising market sizes |
| Strong | Logistics / Supply Chain | World Bank logistics index, Freightos Baltic Index, 3PL reports, FedEx/UPS earnings calls | Shipping volume trends, freight rates, digitization adoption, last-mile economics |
| Strong | EdTech / Education | UNESCO, NCES, Coursera/LinkedIn Learning annual reports, HolonIQ data | Enrollment data, e-learning adoption rates, market sizes, top platform share |
| Moderate | Manufacturing / Industry 4.0 | McKinsey public research, Deloitte manufacturing studies, industry associations | Macro adoption trends; unit economics typically behind paywall |
| Moderate | Real Estate / PropTech | NAR reports, Zillow Research, CBRE/JLL public data | Transaction volumes, price trends, platform adoption; quality varies significantly by geography |
| Moderate | Insurance | OECD Insurance Statistics, AM Best public data, regulatory filings | Premium volumes, loss ratios at market level; pricing and actuarial data rarely public |
| Thin | Niche B2B Industrial | Limited analyst coverage; trade press exists but sparse in English | Directional framing with ranges; hard numbers require paid sources |
| Thin | Emerging Market Niches | Regional sources exist; analyst coverage in English is limited | General narrative with ranges; country-level data requires local-language sources |
bmad-domain-research Domain Taxonomy — 25 Domains
The agent uses live web search to research any domain. These 25 domains have the richest publicly available material and produce the most actionable output.
Regulated Industries — Highest Research Depth (13 domains)
Regulated · 01
Healthcare / MedTech
- patient, clinical, EHR, EMR, PHI, telehealth
- diagnosis, prescription, hospital, medical device
- lab, radiology, doctor, nurse, health record
Regulated · 02
Financial Services / Fintech
- banking, lending, investment, trading, portfolio
- brokerage, wealth management, credit, loan, mortgage
- fund, financial advisor, GAAP, IFRS
Regulated · 03
Payment / Transactions
- payment, checkout, POS, card processing, wallet
- transaction, settlement, remittance, money transfer
- billing, invoice, subscription billing
Regulated · 04
Insurance
- premium, policy, claim, underwriting, actuarial
- coverage, risk assessment, insurer, broker
- reinsurance
Regulated · 05
HR / Payroll / Employment
- payroll, employee, benefits, compensation, timesheet
- onboarding, offboarding, recruitment, workforce
- contractor, W-2, tax withholding, PTO, FMLA
Regulated · 06
Education / EdTech
- student, learning, curriculum, assessment, LMS
- academic, classroom, tuition, enrollment, course
- grade, transcript, training, certification
Regulated · 07
Government / Public Sector
- government, federal, municipal, citizen service
- public procurement, grant, compliance audit
- regulatory filing, e-government
Regulated · 08
Legal / LegalTech
- contract, litigation, legal document, e-signature
- compliance, court, attorney, law firm, case management
- discovery, IP, patent
Regulated · 09
Telecommunications / Messaging
- telco, SMS, carrier, network, broadband, mobile operator
- MVNO, messaging, push notification
- marketing email, opt-in, opt-out
Regulated · 10
Real Estate / PropTech
- property, mortgage, rental, tenant, lease
- real estate agent, title, escrow, HOA
- construction permit, building code
Regulated · 11
Energy / Utilities / CleanTech
- energy, power grid, renewable, solar, electricity
- utility, emissions, carbon, grid management
- metering, smart grid
Regulated · 12
Food / Beverage / FoodTech
- restaurant, food delivery, recipe, nutrition, FMCG
- hospitality, food safety, supply chain traceability
- allergen, FDA food
Regulated · 13
Agriculture / AgriTech
- crop, farm, irrigation, precision agriculture, soil
- livestock, harvest, supply chain, agri-input, farmer
High-Growth Tech Industries — Rich Analyst Coverage (9 domains)
Tech · 01
E-commerce / Retail
- marketplace, cart, checkout, merchant, storefront
- inventory, fulfillment, D2C, dropship, SKU
- returns, logistics
Tech · 02
SaaS / B2B Software
- subscription, MRR, ARR, churn, onboarding
- enterprise software, seat-based, PLG, product-led growth
- SLA, multi-tenant
Tech · 03
Consumer Apps / Mobile
- app store, DAU, MAU, retention, push notification
- freemium, in-app purchase, mobile-first
- social feature, feed, engagement loop
Tech · 04
Logistics / Supply Chain
- shipping, warehouse, fleet, fulfillment center
- last mile, freight, 3PL, carrier, routing
- inventory management, WMS, TMS
Tech · 05
Manufacturing / Industry 4.0
- factory, production, quality control, IoT sensor, ERP
- assembly line, OEE, defect tracking, MES
- preventive maintenance
Tech · 06
Travel / Hospitality
- hotel, booking, airline, OTA, travel agent
- vacation rental, tourism, itinerary, GDS
- loyalty program, PMS
Tech · 07
Media / Entertainment
- streaming, content, subscription media, creator
- podcast, gaming, publishing, DRM, rights management
- ad-supported
Tech · 08
Construction / ConTech
- project management, BIM, general contractor
- subcontractor, site safety, permit, RFI
- change order, punch list
Tech · 09
Automotive / Mobility
- vehicle, fleet management, EV, connected car
- ride-sharing, autonomous driving, telematics
- dealership, OEM
Emerging & Specialized (3 domains)
Emerging · 01
Cybersecurity
- threat, vulnerability, penetration testing, SOC, SIEM
- zero trust, identity management, firewall
- incident response, audit log
Emerging · 02
AI / ML Platform
- model training, inference, MLOps, LLM, vector database
- embedding, fine-tuning, dataset, GPU, model serving
Emerging · 03
Non-profit / Social Impact
- donation, volunteer, grant, impact measurement
- NGO, charity, fundraising, beneficiary, program outcomes
Multi-domain products: When a product spans multiple domains (e.g., a healthcare payment platform = Regulated·01 + Regulated·02 + Regulated·03), run one Domain Research session scoped to the primary domain, then explicitly include secondary domains in Step 4 (Regulatory Focus) and Step 5 (Technical Trends).
bmad-technical-research Technology Domain Coverage
Coverage comes from: (1) training knowledge through August 2025; (2) live web search for current adoption data, release notes, and community consensus. The agent flags areas where web search confirmation is particularly critical.
Tier 1 — Expert Depth (15 domains)
Tier 1 · 01
Web Frontend
React, Vue, Angular, Next.js, Nuxt.js, TypeScript, Tailwind CSS, Vite, state management (Redux/Zustand/Pinia), SSR vs CSR vs SSG, Core Web Vitals
Tier 1 · 02
Backend / Server-side
Node.js (Express/Fastify/NestJS), Python (FastAPI/Django/Flask), Java/Spring Boot, Go (Gin/Echo/Fiber), Rust (Actix/Axum) — language trade-offs, app structure, background jobs
Tier 1 · 03
Mobile
React Native (Expo & bare), Flutter/Dart, Swift/SwiftUI/UIKit (iOS), Kotlin/Jetpack Compose (Android) — cross-platform vs native, navigation, state, push notifications
Tier 1 · 04
Cloud Platforms
AWS (IAM/EC2/ECS/EKS/Lambda/S3/RDS/Aurora/DynamoDB/SQS/SNS/CloudFront/CDK), GCP (GKE/Cloud Run/BigQuery/Firebase), Azure (AKS/Functions/Cosmos DB/Service Bus/Azure AD)
Tier 1 · 05
Containers & Orchestration
Docker (multi-stage builds), Kubernetes (pod design/ingress/RBAC/HPA/Helm), ArgoCD/Flux GitOps, Istio/Linkerd service mesh
Tier 1 · 06
Relational Databases
PostgreSQL, MySQL, SQL Server — schema design, indexing (B-tree/GIN/partial), query optimization, migrations (Flyway/Alembic), connection pooling, replication
Tier 1 · 07
NoSQL & Caching
MongoDB (document modeling, aggregation), Redis (data structures/pub-sub/streams/cluster), DynamoDB (single-table design/GSI), Cassandra, Elasticsearch/OpenSearch
Tier 1 · 08
API Design
REST (versioning/pagination/HATEOAS), GraphQL (Apollo/N+1/DataLoader/subscriptions), gRPC (protobuf/streaming), WebSockets, Webhooks (delivery/retry/idempotency)
Tier 1 · 09
Auth & Authorization
OAuth 2.0 flows (auth code/PKCE/client credentials), OpenID Connect, JWT (expiry/rotation/revocation), RBAC/ABAC, SAML 2.0, API key management
Tier 1 · 10
CI/CD & DevOps
GitHub Actions, GitLab CI, CircleCI, Jenkins — pipeline design, blue-green/canary/rolling deployments, feature flags, rollback patterns
Tier 1 · 11
Infrastructure as Code
Terraform (state/modules/workspaces), Pulumi, AWS CDK (L1/L2/L3 constructs), Ansible
Tier 1 · 12
Microservices Patterns
API Gateway (Kong/NGINX), Service Discovery (Consul), Circuit Breaker (Resilience4j), Saga (choreography vs orchestration), CQRS, Outbox, Strangler Fig
Tier 1 · 13
Messaging & Event Streaming
Kafka (topic design/partitioning/consumer groups/exactly-once/Kafka Streams), RabbitMQ (exchange types/DLQ), SQS/SNS, Redis Pub/Sub and Streams
Tier 1 · 14
Observability
Prometheus (PromQL/alerting), Grafana, OpenTelemetry (traces/metrics/logs), Datadog (APM/synthetics), Jaeger/Zipkin, ELK/EFK stack
Tier 1 · 15
Testing
Unit (Jest/Pytest/JUnit/Go testing), Integration (Testcontainers), E2E (Cypress/Playwright/Selenium), Load (k6/Locust/JMeter), Contract (Pact), test pyramid
Tier 2 — Strong Coverage (9 domains)
Tier 2 · 01
Data Engineering
Apache Spark (RDD/DataFrame/streaming), Airflow (DAG design/operators/XCom), dbt (models/tests/macros), Fivetran/Airbyte, Delta Lake/Apache Iceberg/Hudi
Tier 2 · 02
Data Warehousing
Snowflake (virtual warehouses/Time Travel), BigQuery (partitioning/clustering/slots), Redshift (distribution/sort keys), dimensional modeling (Kimball/data vault)
Tier 2 · 03
AI / ML Platform
TensorFlow, PyTorch, scikit-learn, Hugging Face, LangChain/LlamaIndex, OpenAI API (function calling/embeddings), RAG architecture, vector DBs (Pinecone/Weaviate/Chroma/pgvector), MLflow/W&B
Tier 2 · 04
Serverless / Edge
AWS Lambda (cold start mitigation/layers/power tuning), Vercel/Netlify (edge functions/ISR), Cloudflare Workers (KV/R2/Durable Objects), Azure Functions
Tier 2 · 05
Search
Elasticsearch/OpenSearch (index mapping/query DSL/dense vector semantic search), Typesense (faceting/typo tolerance), Algolia (instant search/ranking)
Tier 2 · 06
Payment & Billing
Stripe API (PaymentIntents/subscriptions/webhooks/Connect), Braintree, Adyen, PayPal — PCI DSS patterns, webhook idempotency, subscription lifecycle
Tier 2 · 07
Notifications
Email: SendGrid/Mailgun/SES (deliverability/SPF/DKIM/DMARC); SMS: Twilio; Push: Firebase FCM/APNs — opt-in management, suppression lists
Tier 2 · 08
Identity Platforms
Auth0 (actions/M2M), Okta (workforce/CIAM), AWS Cognito, Firebase Auth, Clerk — SSO, MFA, social login patterns
Tier 2 · 09
Security Engineering
OWASP Top 10, HashiCorp Vault/AWS Secrets Manager, WAF patterns, SAST/DAST (SonarQube/Snyk/ZAP), security headers, TLS configuration
Tier 3 — Adequate for Scope & Trade-off Evaluation (6 domains)
Tier 3 · 01
IoT & Edge
MQTT (QoS levels/broker selection: Mosquitto/HiveMQ/EMQX), AWS IoT Core, Azure IoT Hub, edge vs. cloud processing, FOTA, InfluxDB/TimescaleDB
Tier 3 · 02
Blockchain / Web3
Ethereum architecture, Solidity basics, Hardhat/Foundry, ERC-20/721/1155, gas optimization, ethers.js/web3.js, Layer 2 (Optimism/Arbitrum/Polygon)
Tier 3 · 03
AR / VR / XR
ARKit (iOS), ARCore (Android), WebXR API, Unity for XR — scene anchoring, spatial audio, performance constraints; Unreal Engine basics
Tier 3 · 04
Embedded / Firmware
C/C++ for microcontrollers, Arduino, Raspberry Pi (Linux-based), FreeRTOS, I2C/SPI/UART protocols, power management patterns
Tier 3 · 05
Low-code / No-code
Bubble, Webflow, Retool (internal tools), Airtable as backend, AppSheet — capability boundaries, when to switch to custom code, REST API integration patterns
Tier 3 · 06
Game Development
Unity (MonoBehaviour lifecycle/ScriptableObjects/ECS/DOTS), Godot (node system/GDScript), Unreal Engine — networking models (authoritative server/client prediction), asset pipeline