5 months ago
Remote, AmericasSenior
Responsibilities
- Ship user-facing agent experiences from prototype through production and iteration based on real usage.
- Architect and implement stateful agent systems involving workflows, tool calling, memory, retrieval, and human-in-the-loop processes.
- Build real-time and asynchronous voice features, including speech agents, voice interfaces, audio routing, and safe tool-execution guardrails.
- Build evaluation harnesses with curated test sets, scenario suites, automated scoring, rubric-based graders, prompt/model/version tracking, canary testing, A/B experiments, and safe rollouts.
- Design data and retrieval pipelines covering chunking, enrichment, metadata, hybrid retrieval, reranking, caching, latency optimization, and multi-tenant data isolation.
- Integrate with and extend Django/DRF/ASGI services, asynchronous execution, queues, workflow orchestration, PostgreSQL, pgvector, and Kubernetes deployments.
- Establish observability, reliability, security, and privacy practices for production agent systems, including traces, structured logs, timeouts, retries, circuit breakers, and graceful degradation.
Requirements
- Strong software engineering fundamentals in design, testing, code quality, performance, and security.
- Production experience deploying AI systems for users, beyond notebooks or demos.
- Experience building agentic or LLM-powered systems with memory and tool use.
- Ability to work across APIs, background jobs, data stores, and deployment infrastructure.
- Hands-on experience with an agent framework or equivalent custom implementation, such as LangChain, LangGraph, LlamaIndex, AutoGen, or CrewAI-style multi-agent patterns.
- Strong understanding of retrieval and vector search concepts, including embeddings, indexing, filtering, and evaluation.
- Preferred experience with vector databases or search stacks such as Pinecone, Chroma, Weaviate, Qdrant, or pgvector.
- Preferred experience designing offline and production evaluation systems, human evaluation loops, monitoring, and prompt/model regression testing.
- Preferred experience building voice or real-time systems, including streaming, WebRTC, or speech-to-text/text-to-speech integrations.
- Preferred experience with durable, long-running workflows such as Temporal or similar orchestration engines.
- Preferred familiarity with observability tools such as OpenTelemetry or Datadog.
- Preferred experience shipping multi-tenant SaaS systems with strong privacy boundaries.
- Must be available for overlap with Americas time zones and have reliable high-speed internet.
Benefits
- Remote work arrangement.
- Full-time employment.
- Overlap with Americas time zones required for collaboration.
- Reliable high-speed internet required.