Backend Engineer (Grok Voice Agent)

Mô tả công việc

Role summary

Senior-leaning backend engineer who owns the real-time voice layer behind the Quality Realty Marketplace (QRM) and its re-sellable AI Native Marketplace Platform (ANMP). You build and operate the Grok Voice Agent integration — the WebSocket proxy, the tool-calling framework that lets our agent "Q" execute real platform actions (valuations, CRM, vendor matching, workflow state changes, payments), and the text/chat fallback. This is the single most central engineering role in our AI-native vision: voice is the primary interface, not a feature.

You work closely with the Tech Lead and the Middle Fullstack Developer, and you set the patterns the Junior Backend Developer and Junior Flutter Developer follow when they consume your APIs.

Current engineering phase (May–July 2026)

Production today runs on a WordPress plugin with REST /wp-json/qr/v3 and paying Dutch realtor tenants. New Python/FastAPI services are added alongside using a strangler-fig approach — no big-bang rewrite. The voice track is net-new: you build the FastAPI voice proxy and tool endpoints from the ground up, while respecting the existing API contracts and multi-tenant model (organization_id, roles/capabilities). Web clients use React/Next.js; mobile uses Flutter (LiveKit on the device) — both consume your real-time layer via APIs.

About Quality Realty (QR)

Quality Realty is an early-stage AI-native SaaS startup building the future of real estate transactions in the Netherlands. We are developing Quality Realty Marketplace (QRM). This voice-first, multilingual platform connects realtors, clients, and service vendors (painters, movers, stagers, energy auditors, etc.) throughout the entire home buying, selling, and rental process.

At its core is the AI Native Marketplace Platform (ANMP) — a fully re-sellable, white-label, multi-tenant marketplace operating system. QR is the first vertical (Dutch real estate), designed from day one to power other local-service marketplaces worldwide.

Key responsibilities

• Build and maintain the Grok Voice Agent WebSocket proxy in Python/FastAPI: low-latency, streaming, reconnect-safe, with LiveKit integration points for the Flutter mobile client.
• Design a clean, versioned "tools" API that the voice agent can call to execute real actions — CRM operations, vendor matching, property valuations, workflow/order state changes, and payment flows — with strict input validation and per-tenant authorization on every call.
• Implement authentication and session management for voice sessions: JWT-based auth middleware, tenant context propagation (organization_id, roles), session lifecycle, and a robust text/chat fallback for when voice is unavailable.
• Productionize the first 8–10 core voice commands mapping to real realtor actions, end-to-end (intent → tool call → backend mutation → spoken/visual confirmation).
• Integrate external AI/LLM and real-time services (Grok primary; Anthropic/OpenAI where relevant) and coordinate with the Middle Fullstack Developer on WP-side tool endpoints during transition.
• Handle observability and reliability: structured logging, tracing on the streaming path, error handling, idempotency, and graceful degradation (Cloud Logging, Sentry).
• Write OpenAPI specs, tool-function docs, and developer handoff notes so the Junior Backend, Junior Flutter, and Next.js engineers can consume and extend the voice layer safely.
• Enforce security and privacy on the voice path: no PII in logs/metadata where avoidable, GDPR-aware transcript handling, access control on sensitive tools (payments, payouts).
• Participate in code reviews, follow team standards (ruff/black, mypy, pre-commit where adopted), and partner with the Tech Lead on architecture and scalability

Yêu cầu công việc

Required skills & experience

• 3–5 years of backend engineering, with proven delivery of production services.
• Strong Python (FastAPI preferred; async/await, packaging, testing) — required.
• Hands-on real-time systems experience: WebSockets, streaming, low-latency APIs, connection lifecycle, and backpressure handling.
• Experience integrating AI/LLM APIs (Grok, Anthropic, OpenAI) or voice platforms (LiveKit, Twilio, Vonage, or similar).
• Solid understanding of authentication and session state (JWT/OAuth, secure token handling) and cloud deployment concepts (containers, serverless; GCP a plus).
• Comfortable designing and documenting JSON/REST APIs and OpenAPI specs.
• Working knowledge of relational databases (PostgreSQL preferred) and safe schema/migration practices under supervision of the Tech Lead.
• Good English communication (written and spoken) for documentation, team discussions, and collaboration with the Dutch CTO and Tech Lead.

Nice to have

• Experience building tool-calling / function-calling frameworks for LLM agents.
• Familiarity with LiveKit, WebRTC, or audio streaming pipelines.
• Exposure to multi-tenant SaaS (tenant isolation, RLS, configurable rules).
• Stripe Connect / payments integration awareness (you may expose payment tools to the agent).
• Experience with Google Cloud (Cloud Run, Cloud SQL, Pub/Sub) or similar.
• CI/CD (GitHub Actions / GitLab CI), Docker, monitoring (Sentry).
• Interest in voice UX, multilingual systems (Dutch & English), or real estate / proptech.
• Basic awareness of WordPress REST or PHP (you do not own the plugin, but you align with its API contracts during transition).

Tại sao bạn sẽ yêu thích làm việc tại đây

What we offer

• Ownership of the defining technical differentiator of the platform — the voice-first layer real Dutch realtors use daily.
• Direct collaboration with the Tech Lead and CTO; high autonomy and influence over the AI architecture for QR and future ANMP resellers.
• Clear path to voice/AI lead responsibility as QR scales and the platform is licensed to white-label partners.
• Hybrid environment with flexible hours (HCMC-based core team; sync windows with Europe).
• Competitive full-time salary for a senior/mid-senior level in Vietnam + performance bonuses (details with CTO).
• Hands-on work with Grok Voice Agent, FastAPI, real-time streaming, GCP, and real-world marketplace problems during a live production transition.
• Full-time; Wave 2 hire (from July 2026) — joining as the voice foundation is established, ahead of the broader team ramp.

Reporting line

This role reports to the Tech Lead with a close working relationship to the CTO on AI/voice strategy. Works closely with the Middle Fullstack Developer (WordPress + Python) (WP-side tool endpoints), the Junior Backend Developer (Python services), the Junior Flutter Developer (mobile LiveKit client), the Data Scientist (valuation tools exposed to "Q"), and the UX Designer (voice-to-visual flows). The Next.js web track consumes the same APIs.