AI Engineer (Full Stack)

We’‘re building AI-powered products end to end models, backend services, and the interfaces people actually use. You’‘ll own features across that whole stack rather than sitting in a narrow research or infra lane. This is a hands-on build role: you should be comfortable shipping production code daily and using AI coding tools as a core part of how you work, not as a novelty., * Design, build, and ship full-stack features backed by LLMs - from prompt and context design through API layer, data model, and frontend.

• Build and productionize applications on the Gemini model family via Vertex AI / Gemini API: structured outputs, function calling, tool use, grounding, multimodal inputs, and long-context patterns.
• Implement RAG and agentic workflows retrieval, embeddings, vector search, chunking strategy, tool orchestration, and multi-step agent loops.
• Deploy and operate services on Google Cloud Platform: Cloud Run, GKE, Cloud Functions, BigQuery, Firestore/Cloud SQL, Pub/Sub, Cloud Storage, Secret Manager.
• Use AI coding agents (Codex, Gemini CLI / Code Assist, Claude Code, Cursor, or similar) as part of your daily workflow, and help the team establish good practice around them review discipline, test coverage, and where agents should and shouldn’‘t be trusted.
• Build evaluation and observability for AI features: eval sets, regression testing on prompt changes, latency and token-cost tracking, tracing, and guardrails against hallucination and prompt injection.
• Own quality end to end tests, CI/CD, monitoring, on-call for what you build.
• Work directly with product stakeholders to turn loose problem statements into scoped, shippable work., * Multi-agent orchestration frameworks (LangGraph, ADK, CrewAI, Genkit).
• Fine-tuning, distillation, or model adaptation on Vertex AI.
• Streaming and realtime interfaces (SSE, WebSockets, voice/multimodal UX).
• Other cloud or model providers AWS Bedrock, Azure OpenAI, Anthropic API, open-weight model serving.
• Data engineering with BigQuery / Dataflow.
• Open-source contributions or public technical writing.

• 3-5 years of professional software engineering, with at least 1 year building production LLM-backed features (prototypes and demos alone won’‘t clear the bar).
• Full-stack capability strong backend (Python and/or TypeScript/Node; Go a plus) plus working frontend competence (React or equivalent). You should be able to take a feature from database to UI without a handoff.
• Google AI stack hands-on with Gemini models via Vertex AI or the Gemini API in a production or near-production setting.
• Google Cloud Platform real deployment and operational experience, not just familiarity. Comfortable with IAM, networking basics, cost control, and infrastructure-as-code (Terraform preferred).
• AI-assisted development demonstrated, fluent use of agentic coding tools (Codex, Gemini CLI, Claude Code, Cursor, Copilot Workspace or similar) on real codebases.
• Applied LLM engineering prompt and context engineering, function/tool calling, RAG, embeddings and vector search, and structured output handling.
• Solid fundamentals: Git, API design (REST/gRPC), relational and document data modeling, containers, CI/CD, automated testing.
• Strong written communication this is a distributed team and decisions live in docs and PRs.