Kora is a Toronto-based, VC-backed startup building an intelligent assistant for busy families. Our product combines a mobile app with a dedicated smart home display to help families manage schedules, chores, shopping, meals, and everyday responsibilities.
We are building an assistant that understands family context, remembers what matters, anticipates what needs attention, and takes action to help families get things done.
We are a small, hands‑on team. You will work directly with the founders to build Kora’s core intelligence and shape how the assistant behaves in everyday life.
The role
We are looking for an AI engineer with strong backend development skills to build and improve Kora’s core assistant capabilities.
You will work on the full agent loop: understanding a request, gathering relevant context, deciding what to do, calling tools, checking results, and following through. Your work will include prompts, orchestration, memory, integrations, evaluations, and the services that run these workflows.
You should enjoy turning emerging AI capabilities into useful product features and making them dependable enough for real users.
What you’ll build
- Agent reasoning and execution: The prompts, tool definitions, control flow, and execution harness that enable Kora to complete tasks across multiple steps.
- Family context and memory: Systems that help Kora understand people, preferences, schedules, responsibilities, and ongoing commitments.
- Proactive assistance: Workflows that detect changes, conflicts, preparation needs, and deadlines, then decide when and how to help.
- Tools and integrations: Connections to calendars, tasks, shopping lists, email, and other services that let the assistant take action.
- Conversational experiences: Support for requests through chat, voice, images, and forwarded information.
- Reliable backend infrastructure: APIs, background jobs, event processing, and persistent workflow state.
- Evaluation and observability: Tools to understand agent behavior, measure quality, investigate failures, and improve performance.
What you’ll do
- Own AI features from initial design and prototype through production deployment and iteration.
- Develop and refine system prompts, context assembly, structured outputs, and tool-calling behavior.
- Build Python backend services that connect models, application data, and external integrations.
- Design workflows that can handle missing information, interruptions, retries, and partial failures.
- Implement permissions and confirmation flows so the assistant acts appropriately on a family’s behalf.
- Build evaluations using realistic scenarios and production failure cases.
- Investigate agent traces and user feedback to identify why a workflow succeeded or failed.
- Improve response quality, latency, and inference cost.
- Collaborate with mobile and device engineers to deliver a consistent assistant experience.
- Help decide when to use an LLM, a deterministic workflow, or a combination of both.
What we’re looking for
- Strong Python skills and experience building backend services, APIs, and database-backed applications.
- Hands‑on experience building an LLM‑powered application beyond a basic tutorial.
- Practical understanding of prompting, tool calling, structured outputs, context management, and retrieval.
- Experience shipping software and diagnosing problems in production.
- Ability to reason about asynchronous tasks, application state, error handling, and data consistency.
- A systematic approach to testing AI behavior and evaluating whether changes actually improve the product.
- Product judgment: you care about whether the assistant solves the user’s problem and earns their trust.
- Regular use of AI development tools, with ownership of the quality and correctness of the resulting code.
- Comfort working independently in a small team, clarifying ambiguous problems, and delivering complete features.
Helpful experience
- Production AI agents or workflows that execute actions through external tools.
- Agent memory, context engineering, retrieval systems, or knowledge graphs.
- Durable workflows, job queues, scheduling, and event‑driven systems.
- PostgreSQL, Firebase, GCP, or similar infrastructure.
- OAuth and integrations with calendar, email, or productivity services.
- Voice assistants, real‑time streaming, or multimodal applications.
- Agent evaluations, tracing, monitoring, and regression testing.
- Handling untrusted inputs, prompt injection, and permissions in AI applications.
We value strong engineering fundamentals, practical judgment, and evidence of building useful things. You do not need experience with every framework or technology listed.
- Build the core intelligence of a consumer AI product.
- Work directly with the founders and influence architecture and product behavior.
- Tackle real problems involving memory, reasoning, proactive assistance, and action.
- Connect AI capabilities to mobile software and a physical product in people’s homes.
- Take substantial ownership and grow with an early‑stage, VC‑backed company.
How we work: AI-native from day one
Kora is a fully AI-native company. AI coding agents are central to how we research, build, debug, test, and ship software. We expect every engineer to work effectively with tools such as Cursor, Claude Code, Codex, or equivalent platforms.
You must be able to:
- Use agentic coding workflows to deliver complete features, from planning and implementation through testing and review.
- Give coding agents clear context, break complex problems into manageable tasks, and guide their work across an existing codebase.
- Use AI tools to investigate bugs, explore unfamiliar code, refactor systems, and automate repetitive development tasks.
- Critically review generated code, verify its behavior, and take responsibility for correctness, security, and maintainability.
- Continuously improve your development workflow as tools and capabilities evolve.
We look for practical experience using coding agents to ship working software, alongside the engineering judgment to recognize when their output needs correction.