Lydia AI - ORCALydia AI is building ORCA, an agentic AI platform that helps businesses research, analyze, and present complex information using AI-powered workflows.ORCA helps enterprises scale expert‐led work by encoding how complex work gets done. Our platform turns source materials, business rules, human judgment, and review processes into repeatable agentic workflows that produce completed work products such as decks, reports, evidence packs, plans, campaigns, and other enterprise deliverables.The OpportunityWe are looking for a senior engineer to help scale the core runtime behind these workflows.This is not a prompt engineering role. This is an engineering‐heavy role for someone who can design, build, debug, and scale production systems where LLMs, tools, retrieval, workflow state, databases, and human review need to work reliably together.The RoleAs a Senior AI Systems Engineer, you will help build and scale ORCA's agentic workflow runtime. You will work across backend services, agent orchestration, retrieval, workflow state, persistence, service boundaries, and production reliability.We are looking for someone roughly 10+ years into their engineering career, with meaningful experience building and scaling production systems.Key ResponsibilitiesBuild and scale production‐grade Python backend servicesImprove ORCA's agent runtime, gateway, connector, and persistence architectureDesign reliable agent workflows using tool calling, function calling, structured outputs, retries, and fallbacksBuild retrieval and source‐grounding systems for enterprise source materialsOwn PostgreSQL / Supabase schema design, migrations, RLS, query performance, and workflow state persistenceDefine clean API and service boundaries across runtime, connectors,



frontend surfaces, and downstream toolsInstrument agent runs with logs, traces, evals, and regression testsImprove reliability, latency, cost, debuggability, and production recoveryHarden tenant isolation, permission boundaries, auditability, and enterprise securityMentor engineers and raise the technical bar across the platformWhat We're Looking ForYou should have deep experience with:Production backend engineering, ideally in Python, including async execution, task lifecycle, retries, timeouts, cancellation, background workers, and production debuggingAgentic AI systems, including LLM tool use, function calling, prompt routing, multi‐turn state, context handling, retries, fallbacks, and multi‐step workflow executionRetrieval and source grounding, including RAG, embeddings, semantic search, source search, context assembly, document grounding, and retrieval qualityPostgreSQL and data modeling, including schema design, migrations, indexing, query performance, transaction boundaries, and debugging query plansAPI and service architecture, including service contracts, versioning, idempotency, failure handling, and clean boundaries between systemsEvaluation and observability, including traces, logs, eval sets, golden cases, regression testing, and root‐cause analysis for agent failuresEnterprise security, including tenant isolation,



permission boundaries, audit logs, least‐privilege access, and data leakage preventionOperational debugging, including the ability to reason from logs, database state, service behavior, and imperfect production environmentsORCA currently includes some non‐standard runtime components, including macOS / launchd‐managed services, so you should be willing to operate outside standard Linux‐only assumptions when needed.Strong-to-Have ExperienceLangGraph, LangChain, CrewAI, MCP, or similar orchestration frameworksAnthropic, OpenAI, or other tool‐use / function‐calling model APIsMulti‐model routingTelegram Bot API or chat‐based workflow interfacesConnector server designDocument ingestion, structured extraction, or source‐pack processingSupabase RLS at scaleMulti‐tenant SaaS authorizationCI/CD, deployment, monitoring, and infrastructure hygieneFrontend awareness, especially around surfacing workflow state and agent review states to usersYou'll Be a Robust Fit If YouHave scaled production systems before, not just built prototypesAre comfortable owning architecture and implementationCan debug complex failures across services, models, tools, and databasesThink carefully about workflow state, permissions, and data boundariesUnderstand that agentic systems require engineering discipline, not just better promptsCan help a fast‐moving product mature into a reliable enterprise platformIdeal BackgroundWe expect the right candidate to be approximately 10+ years into their engineering career, with experience in backend systems, platform engineering, infrastructure, AI systems, or enterprise SaaS. #J-18808-Ljbffr

📌 Senior Ai Systems Engineer (Toronto)
🏢 Knowtions
📍 Toronto