F
Artificial Intelligence Engineer
Found People Inc.
2 hours ago
Full-time
On-site
Applied AI Engineer – Agentic Systems
Confidential AI Company | $100M-Backed | Remote – Canada & US
About the Opportunity
Building an AI demo is easy. Building an agent people trust to actually do the work is the hard part.
We're partnering with a well-backed AI company building the next generation of proactive, AI-native applications for everyday life.
Billions of people rely on traditional applications for email, calendars, notes and tasks. Our client's mission is to fundamentally rethink these experiences by building intelligent applications that understand user intent, maintain context over time and complete real-world tasks with minimal prompting.
The challenge isn't simply making AI smarter. It's making AI
reliable enough to take action on behalf of users. As an
Applied AI Engineer , you'll build the systems that turn frontier language models into dependable, production-ready agents while contributing to the applications and backend infrastructure that bring those capabilities to users. Your work will span
agent orchestration, persistent memory, tool use, evaluation, backend engineering and full-stack product development . This is a highly hands-on role within a small, talented engineering team. You'll work closely with ML, product and application engineers, with meaningful ownership over architecture, implementation and technical decisions.
Experience, qualification, and soft skills, have you got everything required to succeed in this opportunity Find out below.
What You'll Own Agent Architecture & Orchestration Design and build AI agents that plan, reason and execute complex, multi-step tasks. Develop orchestration systems connecting models, tools, APIs and external services. Build long-running workflows that maintain state, preserve user intent and operate reliably across sessions. Implement tool execution with retries, fallbacks, error handling and recovery from interruptions. Prevent duplicate, unintended or unauthorized actions through appropriate execution controls. Develop human-in-the-loop approvals and escalation mechanisms for sensitive actions. Evaluate single-agent and multi-agent approaches, balancing reliability, complexity, performance and cost.
Memory & Context Management Design systems for short-term working memory and long-term persistent context. Build retrieval and grounding mechanisms connecting agents to relevant user information and application data. Develop context management strategies that improve performance while controlling latency and cost. Address challenges involving stale information, conflicting context, memory updates and user privacy. Enable agents to understand user preferences, prior interactions and evolving intent without repetitive prompting.
LLM Integration & Optimization Integrate frontier and open-source language models into production applications. Build reliable model interactions using structured outputs, tool schemas and prompt engineering. Evaluate model performance based on task completion, reliability, latency and cost. Implement model routing, caching and context optimization strategies. Partner with ML engineers on model adaptation and fine-tuning when appropriate. Improve agent performance through production data, evaluation and user feedback.
Evaluation, Reliability & Observability Build evaluation frameworks measuring end-to-end task completion rather than individual model responses. Develop automated tests and benchmarks for agent behaviour, tool execution and workflow reliability. Implement tracing, monitoring and debugging for complex AI workflows. Identify and reduce hallucinations, incorrect tool calls and cascading failures. Design safeguards around sensitive actions, permissions, external integrations and user data. Establish measurable standards for reliability, latency, cost and successful task execution.
Full-Stack Product Engineering Build backend services, APIs and application features that bring AI capabilities into email, calendar, notes and productivity applications. Develop user-facing features using modern frontend technologies such as React, Next.js and TypeScript. Design scalable APIs, databases and integrations supporting intelligent application workflows. Collaborate with product and design teams to create intuitive experiences around complex AI capabilities. Own features and systems from initial concept through development, deployment and ongoing improvement. Make pragmatic engineering decisions balancing speed, reliability, maintainability and user experience.
What We're Looking For We're looking for
strong software engineers who have moved beyond experimenting with LLMs and understand what it takes to build dependable AI-powered products. The ideal candidate will bring: Strong software engineering fundamentals
and experience building, shipping and supporting production applications. Strong Python experience , including backend services, APIs and asynchronous workflows. Hands-on experience building and deploying LLM-powered applications , ideally involving agents, orchestration, tool calling or multi-step workflows. Experience across the full stack, with modern frontend technologies such as
React, Next.js, JavaScript or TypeScript . Strong backend engineering experience, including APIs, databases and scalable application services. Understanding of state management, distributed or asynchronous systems and third-party integrations. Familiarity with LLM evaluation, observability, debugging and production reliability. The ability to independently navigate ambiguous problems, make technical decisions and take ownership from concept through production. Strong product instincts, curiosity and an interest in building applications that people use every day. Comfort working in a small, fast-moving team where engineers are expected to contribute across traditional boundaries.
We're particularly interested in engineers who can explain not only what they've built, but what failed, what they learned and how they made the system more reliable. Strong Assets Experience building
production AI agents
involving multi-step tool use, persistent memory, orchestration or autonomous task execution. Experience with LangGraph, LangChain, AutoGen or comparable custom agent architectures. Experience developing RAG systems, semantic search, embeddings or vector retrieval. Familiarity with model evaluation frameworks, agent tracing and structured testing. Experience with OpenAI, Anthropic or open-source LLMs. Knowledge of model routing, inference optimization, fine-tuning or PyTorch. Experience with event-driven architectures, message queues, persistent workflows or distributed task execution. Familiarity with MCP, tool permissions, structured outputs and secure execution environments. Experience working in early-stage startups or building AI products from zero to one. You don't need experience with every framework. We care more about
strong engineering fundamentals, genuine hands-on AI experience and the ability to turn complex technology into reliable products.
Technology Environment Languages:
Python, TypeScript, JavaScript Frontend:
React, Next.js Backend:
Python, Node.js, APIs and asynchronous services AI/ML:
OpenAI, Anthropic, open-source LLMs, PyTorch Agent Systems:
Tool calling, orchestration, memory, planning and execution Retrieval:
RAG, embeddings, semantic search and context management Data:
SQL and NoSQL databases Infrastructure:
Docker, Kubernetes and cloud platforms Reliability:
Evaluation pipelines, observability, tracing, testing and monitoring
What Success Looks Like AI agents reliably complete meaningful, multi-step tasks in production. Long-running workflows maintain context and recover from interruptions or failures. Tool execution is safe, observable and accurate. AI capabilities integrate smoothly into intuitive, responsive applications. Agent performance improves through measurable evaluations and real-world feedback. Hallucinations, incorrect actions and workflow failures are systematically reduced. Latency and operating costs improve without compromising user experience.
How We Work We're a small, high-talent-density engineering team where everyone has broad ownership and significant influence over the product. Engineers are expected to think independently, make sound technical decisions and stay close to the code. We move quickly, experiment often and learn from real-world usage. But we also understand that
when AI systems take action on behalf of users, reliability isn't optional. We value intellectual curiosity, technical depth, simplicity and a willingness to challenge assumptions. We care less about impressive demos and more about building intelligent systems that consistently deliver real results.
Interview Process Our client values transparency, efficiency and practical engineering ability. The interview process consists of approximately three to four interviews with technical team members, conducted virtually and/or onsite. Candidates can expect a focused process and timely decisions. xsgimln If you're excited about building AI agents that move beyond conversation and reliably get things done, we'd love to hear from you. Apply confidentially through Found People.
reliable enough to take action on behalf of users. As an
Applied AI Engineer , you'll build the systems that turn frontier language models into dependable, production-ready agents while contributing to the applications and backend infrastructure that bring those capabilities to users. Your work will span
agent orchestration, persistent memory, tool use, evaluation, backend engineering and full-stack product development . This is a highly hands-on role within a small, talented engineering team. You'll work closely with ML, product and application engineers, with meaningful ownership over architecture, implementation and technical decisions.
Experience, qualification, and soft skills, have you got everything required to succeed in this opportunity Find out below.
What You'll Own Agent Architecture & Orchestration Design and build AI agents that plan, reason and execute complex, multi-step tasks. Develop orchestration systems connecting models, tools, APIs and external services. Build long-running workflows that maintain state, preserve user intent and operate reliably across sessions. Implement tool execution with retries, fallbacks, error handling and recovery from interruptions. Prevent duplicate, unintended or unauthorized actions through appropriate execution controls. Develop human-in-the-loop approvals and escalation mechanisms for sensitive actions. Evaluate single-agent and multi-agent approaches, balancing reliability, complexity, performance and cost.
Memory & Context Management Design systems for short-term working memory and long-term persistent context. Build retrieval and grounding mechanisms connecting agents to relevant user information and application data. Develop context management strategies that improve performance while controlling latency and cost. Address challenges involving stale information, conflicting context, memory updates and user privacy. Enable agents to understand user preferences, prior interactions and evolving intent without repetitive prompting.
LLM Integration & Optimization Integrate frontier and open-source language models into production applications. Build reliable model interactions using structured outputs, tool schemas and prompt engineering. Evaluate model performance based on task completion, reliability, latency and cost. Implement model routing, caching and context optimization strategies. Partner with ML engineers on model adaptation and fine-tuning when appropriate. Improve agent performance through production data, evaluation and user feedback.
Evaluation, Reliability & Observability Build evaluation frameworks measuring end-to-end task completion rather than individual model responses. Develop automated tests and benchmarks for agent behaviour, tool execution and workflow reliability. Implement tracing, monitoring and debugging for complex AI workflows. Identify and reduce hallucinations, incorrect tool calls and cascading failures. Design safeguards around sensitive actions, permissions, external integrations and user data. Establish measurable standards for reliability, latency, cost and successful task execution.
Full-Stack Product Engineering Build backend services, APIs and application features that bring AI capabilities into email, calendar, notes and productivity applications. Develop user-facing features using modern frontend technologies such as React, Next.js and TypeScript. Design scalable APIs, databases and integrations supporting intelligent application workflows. Collaborate with product and design teams to create intuitive experiences around complex AI capabilities. Own features and systems from initial concept through development, deployment and ongoing improvement. Make pragmatic engineering decisions balancing speed, reliability, maintainability and user experience.
What We're Looking For We're looking for
strong software engineers who have moved beyond experimenting with LLMs and understand what it takes to build dependable AI-powered products. The ideal candidate will bring: Strong software engineering fundamentals
and experience building, shipping and supporting production applications. Strong Python experience , including backend services, APIs and asynchronous workflows. Hands-on experience building and deploying LLM-powered applications , ideally involving agents, orchestration, tool calling or multi-step workflows. Experience across the full stack, with modern frontend technologies such as
React, Next.js, JavaScript or TypeScript . Strong backend engineering experience, including APIs, databases and scalable application services. Understanding of state management, distributed or asynchronous systems and third-party integrations. Familiarity with LLM evaluation, observability, debugging and production reliability. The ability to independently navigate ambiguous problems, make technical decisions and take ownership from concept through production. Strong product instincts, curiosity and an interest in building applications that people use every day. Comfort working in a small, fast-moving team where engineers are expected to contribute across traditional boundaries.
We're particularly interested in engineers who can explain not only what they've built, but what failed, what they learned and how they made the system more reliable. Strong Assets Experience building
production AI agents
involving multi-step tool use, persistent memory, orchestration or autonomous task execution. Experience with LangGraph, LangChain, AutoGen or comparable custom agent architectures. Experience developing RAG systems, semantic search, embeddings or vector retrieval. Familiarity with model evaluation frameworks, agent tracing and structured testing. Experience with OpenAI, Anthropic or open-source LLMs. Knowledge of model routing, inference optimization, fine-tuning or PyTorch. Experience with event-driven architectures, message queues, persistent workflows or distributed task execution. Familiarity with MCP, tool permissions, structured outputs and secure execution environments. Experience working in early-stage startups or building AI products from zero to one. You don't need experience with every framework. We care more about
strong engineering fundamentals, genuine hands-on AI experience and the ability to turn complex technology into reliable products.
Technology Environment Languages:
Python, TypeScript, JavaScript Frontend:
React, Next.js Backend:
Python, Node.js, APIs and asynchronous services AI/ML:
OpenAI, Anthropic, open-source LLMs, PyTorch Agent Systems:
Tool calling, orchestration, memory, planning and execution Retrieval:
RAG, embeddings, semantic search and context management Data:
SQL and NoSQL databases Infrastructure:
Docker, Kubernetes and cloud platforms Reliability:
Evaluation pipelines, observability, tracing, testing and monitoring
What Success Looks Like AI agents reliably complete meaningful, multi-step tasks in production. Long-running workflows maintain context and recover from interruptions or failures. Tool execution is safe, observable and accurate. AI capabilities integrate smoothly into intuitive, responsive applications. Agent performance improves through measurable evaluations and real-world feedback. Hallucinations, incorrect actions and workflow failures are systematically reduced. Latency and operating costs improve without compromising user experience.
How We Work We're a small, high-talent-density engineering team where everyone has broad ownership and significant influence over the product. Engineers are expected to think independently, make sound technical decisions and stay close to the code. We move quickly, experiment often and learn from real-world usage. But we also understand that
when AI systems take action on behalf of users, reliability isn't optional. We value intellectual curiosity, technical depth, simplicity and a willingness to challenge assumptions. We care less about impressive demos and more about building intelligent systems that consistently deliver real results.
Interview Process Our client values transparency, efficiency and practical engineering ability. The interview process consists of approximately three to four interviews with technical team members, conducted virtually and/or onsite. Candidates can expect a focused process and timely decisions. xsgimln If you're excited about building AI agents that move beyond conversation and reliably get things done, we'd love to hear from you. Apply confidentially through Found People.