Fullstack AI Software Engineer - (Agentic Systems, Fullstack Builder)

Fullstack AI Software Engineer - (Agentic Systems, Fullstack Builder)

Explicitly requires vibe coding skills—heavy use of AI-assisted coding tools and building agentic/LLM workflows and RAG pipelines.

About the Role

Build end-to-end AI-first wellbeing intelligence at Sonder by owning the full stack from data models and Snowflake transformations to React dashboards and agentic LLM workflows. Deliver production-ready APIs and autonomous AI pipelines that turn safety data into actionable insights for customers.

Role

Sonder is hiring a Fullstack AI Software Engineer on the Customer Insights and Measurement team to own the end-to-end journey from raw data to customer-facing insights. The role covers front-end dashboards, back-end services, data pipelines, and architecting autonomous, agentic AI workflows.

Key Responsibilities

• Full stack development and end-to-end ownership: design and implement database schemas, Snowflake models, back-end services, and React/Next.js UIs.
• API architecture: build robust back-end services (Python or Node.js) to support data ingestion and high-concurrency reporting.
• AI & agentic systems: design and ship autonomous AI workflows using frameworks like LangGraph or CrewAI and implement multi-step agentic loops and tool-calling patterns.
• RAG & context engineering: build and optimize retrieval-augmented-generation pipelines, chunking strategies, and metadata filtering.
• MCP integration: build and deploy Model Context Protocol servers to connect internal data silos with LLM reasoning engines.
• AI observability: implement evaluation and monitoring frameworks (for example LangSmith, Arize Phoenix) to track quality, latency, and cost in production.
• Data pipelines & reporting: design and optimize Snowflake tables/views, automate insight synthesis with LLMs, and support customer-facing dashboards.

Requirements

• Strong full-stack experience with React/Next.js and backend development in Python (preferred) or Node.js.
• Solid SQL skills and experience with data warehousing; Snowflake experience strongly preferred.
• Experience with agentic systems and understanding of multi-step agentic loops, tool-calling, and structured outputs.
• Production mindset: experience with LLM evaluation, prompt/version control, error handling, and operating systems in production.
• Heavy, practical use of AI-assisted coding tools (e.g., Cursor, ClaudeCode, Copilot) and shipping AI-enabled features to production.

Nice-to-Have

• Experience with vector databases and embedding search (Pinecone, pgvector).
• Experience with dbt for data transformations.
• Familiarity with orchestration frameworks (LangChain, LangGraph) and a portfolio of AI-first side projects or open-source contributions.

Working Arrangement & Checks

• Hybrid and flexible working (office presence expected at times).
• Due to the industry, all employees must complete a National Police Check and Working With Children Check.
• Free Sonder app access for you and your family.
• 2 days study leave.
• Paid parental leave (12 weeks primary, 4 weeks secondary) and return-to-work support.
• Employee Share Option Plan (ESOP).
• Novated leasing and salary packaging options.
• Weekly team perks (free coffee Tuesdays, catered lunch Thursdays) and a dog-friendly office.
• Volunteer leave: 2 days/year.
• Regular social events and team bonding opportunities.

Skills

Full Stack Development, API Architecture, Data Modeling, Data Warehousing, SQL, Agentic Systems, Design, RAG & Context Engineering, LLM Evaluation and Observability, Production Mindset, Prompt/Context Engineering, End-to-end Ownership, AI-assisted Development

Experience Level

Senior