Role Overview You will be the architect, strategist, and cultural driver of Vibecoderz’ engineering DNA. This isn’t a CTO role that floats above the codebase - this is a hands-on, deeply technical position where you’ll set the vision and lead the execution of our multi-agent, AI-first, serverless platform. From low-level system design (Firestore schemas, Pub/Sub orchestration, Redis caches, Neo4j Developer Graph) to pixel-perfect UX outcomes (latency budgets, real-time TutorAgent flows, Vibe Browser optimizations), you will own the technical backbone of Vibecoderz. You’ll manage and mentor a 7-person founding engineering team (PM, FE, BE, Full Stack, AI, Prompt, DevOps, QA) while also representing Vibecoderz’ technology to investors, partners, and the wider developer ecosystem. You will set the gold standard in culture, problem-solving, and engineering excellence. This role is for a true achiever - someone who loves solving hard problems, who thrives under scale and complexity, and who is obsessed with user experience and performance.
Key Responsibilities
- Technical Vision & Architecture
- Own the end-to-end architecture of Vibecoderz’ AI learning platform.
- Design and enforce latency/performance budgets across FE, BE, and AI services (<100ms APIs, <200ms UI interactions).
- Define best practices for multi-agent orchestration, real-time workflows, and memory management.
- System Design & Problem Solving
- Lead low-level system design for all critical components (TutorAgent, Developer Graph, Vibe Browser, Artifact Engine).
- Solve complex scaling challenges (100K+ concurrent users, 99.9% uptime, secure sandboxed mini-app execution).
- Engineering Leadership
- Directly manage the 7-person founding engineering team, instilling a culture of velocity + rigor.
- Define coding standards, review processes, and dev culture - “every line of code matters.”
- Product & UX Obsession
- Work with the PM and design teams to ensure AI-first workflows feel magical, seamless, and reliable.
- Obsess over latency, optimization, and polish as first-class product features.
- Infrastructure & DevOps Strategy
- Oversee CI/CD, observability, and cost optimization across GCP.
- Ensure reproducible infra via Terraform and full traceability via OpenTelemetry.
- AI Leadership
- Guide integration of Gemini, LangSmith, and ADK into agent orchestration pipelines.
- Mentor Prompt Engineers and AI Developers to maximize LLM accuracy, consistency, and safety.
- Culture & Mentorship
- Build a developer-first culture where problem solving, ownership, and craftsmanship are core values.
- Mentor engineers across FE, BE, AI, DevOps, and QA to grow into world-class leaders.
- Stakeholder Communication
- Partner with CEO on fundraising, investor pitches, and due diligence.
- Represent Vibecoderz at conferences, meetups, and community events.
- Security & Governance
- Ensure enterprise-grade security (IAM, guardrails, sandboxing, prompt injection defense).
- Build governance frameworks for safe, compliant AI-driven learning.
- Innovation & Future-Proofing
- Constantly evaluate emerging frameworks (CrewAI, AutoGen, LangGraph, MCP).
- Future-proof Vibecoderz by balancing bleeding-edge experimentation with production stability.
Success Metrics
90 Days (Probation):
- Low-level system design finalized, covering FE, BE, AI, infra, and QA components.
- Core engineering processes in place (code reviews, sprints in Linear, docs in Notion).
- First AI Tutor “Text → Course” → “Mini-App” flow live in production with <200ms latency budget.
12 Months:
- 100K+ MAUs supported on stable infra with 99.9% uptime.
- Vibecoderz platform recognized by developers for speed, polish, and reliability.
- Founding engineering team grown from 7 → 20 with culture intact.
- Investor + community recognition of Vibecoderz’ tech leadership.
Must-Haves
- 15+ years in product engineering, with at least 5+ years in technical leadership roles.
- Proven track record of building high-scale, real-time SaaS platforms.
- Deep system design expertise: distributed systems, real-time infra, low-latency APIs.
- Mastery of modern AI-first stacks: Next.js, FastAPI, GCP, Pub/Sub, Redis, Neo4j, LangChain/ADK.
- Product obsession: demonstrated history of building delightful, user-first experiences.
Nice-to-Haves
- Experience leading at an AI startup.
- Contributions to open-source in AI, infra, or real-time frameworks.
- Prior founder or early founding engineer experience.
- Recognized speaker, writer, or thought leader in developer/AI ecosystems.
Tech Stack Visibility Frontend & User Experience
- Frameworks: Next.js (App Router, SSR/CSR hybrid), React, TypeScript
- UI Components: Shadcn UI, Tailwind, Lucide icons, Inter/Geist typography
- Realtime: WebSockets (Socket.IO) for voice transcription, presence, collaboration
- Motion & Interactivity: Framer Motion for micro-animations
- Auth: Firebase Authentication (Google, GitHub OAuth, email/password)
- Client-Side Specialties: Agentic “Vibe Browser” interface rendering Browser-Use outputs
Backend & API Layer
- Framework: FastAPI (Python, async I/O)
- Orchestration: Google Cloud Run (serverless containers)
- Gateway: Google Cloud API Gateway for secure routing & rate limiting
- Background Jobs: Cloud Tasks & Cloud Workflows for async processing (certificates, video generation, Learner Graph updates)
AI & Agentic Layer
- Frameworks: Google Agent Development Kit (ADK), Vertex AI for Agents
- Models: Gemini Pro, Gemini Pro Vision (screen perception), Gemini Live API (voice), Gemini Flash (fast generation)
- Hybrid Model Router: Gemini-first with fallbacks (Claude 3 Opus, others)
- Prompt Ops: LangSmith / Langfuse for prompt versioning, monitoring, and A/B testing
- Agent Communication: Google Cloud Pub/Sub for asynchronous A2A protocols
- Core Agents: TutorAgent, PlannerAgent, StepPlannerAgent, ScreenPerceptionAgent, BrowserUseAgent, ArtifactGenerationAgent, CodeReviewAgent
Data & Memory Systems
- Primary Database: Firestore (user data, sessions, certificates)
- Graph Database: Neo4j Aura for Developer Graph & skill relationships
- Caching Layer: Redis for working session memory, real-time context
- Memory Model: Archival Memory → Working Memory → Real-Time Context (constructed by TutorAgent)
Infrastructure & DevOps
- IaC: Terraform for GCP infra reproducibility
- CI/CD: GitHub Actions for builds, tests, and Cloud Run deployments
- Observability: OpenTelemetry instrumentation, Google Cloud Trace, Grafana dashboards
- Secrets & Security: GCP IAM, Secrets Manager, Google Firewall, VPC service controls
- Disaster Recovery: Automated backups for Firestore, Redis, and Neo4j
Commerce & Notifications
- Billing: Stripe via Firebase Extension for subscriptions (Pro, Teams)
- Transactional Emails: SendGrid via Firebase Extension (welcome, password reset, certificate ready)
Testing & QA
- Frontend Testing: Playwright / Cypress for E2E flows
- Backend Testing: Pytest, Postman/Newman for API validation
- AI Evals: Golden set frameworks + LangSmith/Langfuse eval pipelines
- Load & Stress Testing: Locust / k6 for concurrency validation
Collaboration & Workflow Tools
- Execution & Tasks: Linear (roadmaps, sprints, bug triage)
- Documentation & Knowledge: Notion (PRDs, system design docs, runbooks)
- Code & Repo Management: GitHub (repos, actions, infra as code)
Assessment Objective: Validate ability to architect, code, and optimize the Vibecoderz AI learning platform under real-world constraints. This is a 3-part assessment, to be submitted as a cohesive PoC + documentation package. Part 1 – System Design & Architecture (Strategic)
- Deliver a low-level architecture document for the Text → Course → Mini-App learning flow.
- Must include:
- Service Decomposition: TutorAgent, PlannerAgent, ArtifactGenerator, BrowserUseAgent.
- Memory Design: Archival Memory (Firestore), Developer Graph (Neo4j), Working Memory (Redis).
- Latency Targets: <200ms user-facing latency; identify bottlenecks and trade-offs.
- Scaling Plan: How to handle 100K concurrent learners.
- Security & Guardrails: Defend against prompt injection, unsafe code execution.
Deliverables: 6–8 page design doc with diagrams, trade-offs, and scaling strategies. Part 2 – Coding Challenge (Hands-On) Build a working prototype of a simplified AI Tutor pipeline.
Requirements:
- FastAPI Backend
- /generate-course: Input: "Teach me React Hooks", Output: JSON structured course (outline → lessons → quiz).
- /generate-miniapp: Input: lesson topic, Output: runnable JS code snippet (sandboxed).
- Store outputs in Firestore.
- Agent Orchestration
- Implement at least two collaborating agents (e.g., CurriculumAgent + CodeAgent).
- Use Pub/Sub (mock or real) for async communication.
- Frontend Stub
- Simple Next.js client that:
- Calls the backend endpoints.
- Renders the outline + code snippet.
- Runs the code snippet in a sandboxed iframe.
- CI/CD & Infra
- Deploy backend to Cloud Run.
- IaC with Terraform for Firestore + Pub/Sub setup.
- GitHub Actions workflow to run tests + deploy.
Part 3 – Evaluation & Culture (Leadership)
- Define engineering culture principles you would set for Vibecoderz (max 2 pages).
- Include: code review practices, latency obsession, “dev-first” rituals, and team scaling approach.
Final Deliverables
- Code Repo: GitHub repository with modular, well-documented code.
- Deployed Service: Cloud Run URL with live endpoints.
- Architecture Doc: PDF/Notion link with design diagrams.
- README: Clear explanation of decisions, trade-offs, and infra setup.
- Optional Loom: 5-min walkthrough of PoC + infra.
Evaluation Criteria
- System Design Depth (25%) – Scalability, clarity, trade-offs.
- Coding Execution (30%) – Code quality, modularity, working PoC.
- Performance & Latency Thinking (20%) – Optimizations, budgets, benchmarks.
- Leadership & Culture Fit (15%) – Vision for engineering culture.
- Documentation & Communication (10%) – Clarity of write-ups + repo.