My home turf: 4+ years shipping Flutter apps to both stores.

• Flutter: 8 apps live on stores
• Dart
• Android: Kotlin services, SMS, notifications
• iOS: Swift platform views, App Store releases
• Riverpod: v3, no-codegen discipline
• GetX / Provider / BLoC
• go_router & deep links
• Offline-first (Drift/SQLite, Hive)

Layered Node.js services that one engineer can operate in production.

• Node.js & Express: TypeScript-first
• MongoDB & Mongoose: aggregation pipelines
• PostgreSQL & pgvector
• Redis: caching, rate limits, agent memory
• Python: FastAPI, voice agents
• REST API design
• JWT / RBAC auth
• WebSockets & SSE streaming

LLM systems that survive production: fallbacks, memory, voice.

• OpenAI / Claude / Gemini: multi-provider routing
• RAG & embeddings: pgvector, Qdrant
• Agent memory systems: two-tier STM/LTM design
• Voice AI: LiveKit, SIP telephony
• Speech-to-text: Deepgram, AssemblyAI, Whisper
• Prompt engineering: calibrated confidence, HITL UX
• Vision LLMs: structured extraction + Zod
• Streaming UX: SSE, token-safe rendering

From Firebase serverless to raw EC2, deployed and operated by me.

• AWS: EC2, S3, SQS, SES, ECS
• Firebase: Auth, Firestore, Functions, FCM, App Check
• Docker & Docker Compose
• nginx + PM2
• Vercel / Render / Railway
• CI/CD & release pipelines
• Sentry / Crashlytics

The unglamorous parts that make apps into businesses.

• RevenueCat: trials, entitlements, paywalls
• Razorpay & Stripe
• Analytics: CleverTap, Amplitude, Mixpanel
• Push & retention loops: FCM, deep links, streaks
• App Store / Play releases: signing to review
• Universal / App Links
• i18n: up to 8 languages

Daily drivers.

• Git & GitHub
• VS Code / Android Studio / Xcode
• Figma: design-to-code fidelity
• Postman
• Claude Code / AI-first workflow
• Google Maps Platform