AI Agent Lead

Own Blanket’s core autonomous agents and turn LLM capability into production-grade outcomes through tight context, tooling, evals, and reliability.

Responsibilities

• Deliver agents (and its specialist sub-agents) to production and keep it improving week after week.
• Build agent orchestration: tool calling, routing, retries, state, and safe fallbacks.
• Own context and memory: STM (session context) and LTM (retrieval, summaries, embeddings, structured stores).
• Run the quality loop: datasets, labeling, eval suites, regression gates, and error analysis.
• Integrate and operate: connect to platform APIs, permissions, and internal services, and track accuracy, latency, and cost with tracing and dashboards.

Qualifications

We are open to different backgrounds. You might be a software engineer, ML engineer, data scientist, or data engineer. What matters is that you have already shipped an AI agent or LLM-powered product and you know how to make it reliable.

Must Have

• 5+ years building production systems in software, data, ML, or data science.
• Strong coding ability and comfort shipping to production. Python is preferred.
• Strong SQL and comfort working directly with real production data.
• Proven experience shipping LLM or agentic systems: tool use, structured outputs, guardrails, and validation.
• Experience running evaluation and experimentation: offline evals, online tests, and regressions.
• Daily development with Cursor and Claude Code.

Nice to Have

• Agent frameworks: Strands Agents, LangGraph, PydanticAI, or similar.
• Retrieval and RAG: embeddings, vector DBs, reranking, context construction.
• Orchestrators and agent hierarchies: tool registries, routing policies, state machines.
• Data pipelines (ETL/ELT), data quality checks, analytics workflows.
• Voice agents: speech flows, ASR/TTS, call routing, conversation state.
• MCP ecosystems and typed tool schemas.
• Cloud ops (AWS preferred) and strong observability.

Collaboration

• CTO: strategic direction, architecture tradeoffs, and outcomes.
• Head of Engineering: delivery, engineering standards, reliability, scaling execution.
• Product: scope, UX, success metrics, rollout.
• Customer Success: workflows, feedback loops, real-world failure modes.
• Domain Experts: property management context, edge cases, operational constraints.

The Company

Blanket is building the first Agentic Customer Lifecycle Management Platform for property managers. We serve hundreds of Property Management companies and tens of thousands of homeowners, helping teams automate and execute complex, real-world workflows across sales, retention, and growth.

Stack

• Python (primary)
• Strands Agents with AWS Bedrock AgentCore
• S3, DynamoDB, Redis, Athena, Glue
• REST, gRPC, GraphQL
• Tracing, structured logs, eval dashboards, cost monitoring

Why Join

• We’re booming: strong momentum and real demand from the market.
• Real impact: what you ship directly improves our business and helps PM teams and homeowners solve real problems.
• High autonomy and fast iteration, with real customers using what you build.
• Work across product, data, and engineering, not in a narrow lane.
• A team that’s fun to work with: high standards, low ego, lots of shipping.
• Competitive compensation and flexibility (hybrid-friendly).