Forward Deployed Engineer

We're hiring a Forward Deployed Engineer to build and deploy production AI systems directly with clients, spanning voice agents, agentic workflows, and document pipelines. You'll architect solutions, ship into live environments, and iterate on real-world usage.
What You'll Do
Build bank statement analyzers covering extraction, parsing, categorization, and anomaly detection from unstructured statements. Design and deploy voice AI agents using Pipecat, handling STT/TTS orchestration, turn detection, VAD, and telephony/SIP integration for real-time conversational use cases. Build agentic applications using LangGraph and LangChain, including multi-step reasoning flows, tool-calling agents, state management, and human-in-the-loop checkpoints. Work directly with client teams to scope requirements, prototype rapidly, and deploy into production. Integrate agents with external systems including CRMs, databases, banking APIs, and internal tooling. Debug latency, accuracy, and reliability issues across live deployments, and build evaluation harnesses for agent behavior. Own the full loop from prompt design to production monitoring.
Requirements
Strong Python skills, comfortable with FastAPI or similar async frameworks. Hands-on experience with LangGraph and LangChain for multi-agent, tool-using systems. Proven experience with Pipecat or similar real-time voice AI frameworks in production. Experience building document extraction pipelines, specifically bank statements or unstructured financial data. Experience with vector databases and RAG architectures for grounded retrieval. Cloud deployment experience on AWS, including ECS, RDS, S3, and Bedrock. Background in fintech, fraud prevention, or financial services domains. Experience with multilingual voice systems and SIP telephony trunking. Strong client-facing communication, able to translate ambiguous requirements into shipped systems.