Key skills
TypeScriptSQLAIMLLLMOpenAIAnthropicGeminiRAGAgenticReactExpressWebSocketsAWSTerraformKubernetesDockerHelmS3SQSAPI GatewayPostgreSQL
About this role
Flock is a leading safety technology platform dedicated to crime prevention and security. They are seeking a Software Engineer III to build Night Shift, a conversational AI assistant aimed at aiding investigators in efficiently surfacing critical evidence and closing cases. The role involves designing conversational interfaces, developing backend orchestration, and integrating AI with Flock's existing data platform.
Responsibilities:
- Design and implement the conversational interface for Night Shift
- Build the orchestration backend that manages LLM interactions and tool calling
- Develop integration pipelines connecting AI to Flock's existing data platform and APIs
- Collaborate closely with ML engineers on prompt engineering and agentic workflows
- Shape chat experiences with complex context management
- Partner with platform teams to design new APIs or leverage existing ones
- Solve the reliability challenges of deploying AI in high-stakes investigative workflows
Requirements:
- Love for coding and continuous learning, especially in the rapidly evolving LLM space
- Resourceful problem-solver mindset: excel in ambiguous situations and take initiative to define product direction
- Strong TypeScript / Node / Express skills for web services and API design (REST, SSE, WebSockets for streaming)
- Modern web framework expertise (React / TypeScript preferred), particularly for conversational UI and chat interfaces
- Hands-on LLM experience: OpenAI/Anthropic/Gemini APIs, prompt engineering, streaming responses, and conversation context management
- Familiarity with agentic patterns: function calling, tool use (MCP), and orchestrating multi-step workflows
- API integration skills: consume existing APIs or design new ones to ground AI in investigative data
- Database confidence: PostgreSQL and sophisticated SQL for data retrieval
- Cloud infrastructure basics: Docker, Kubernetes (Helm), AWS services (S3, SQS, API Gateway)
- Product-minded: translate user feedback into technical requirements and make pragmatic tradeoffs
- LLM evaluation tools (LangSmith, Langfuse)
- vector search/RAG
- microservices architecture
- Terraform