SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: CAD 110,000 - 200,000 / annual
Gigi is an agentic operating system for enterprise media buying, with a flagship AI media manager for Amazon DSP. Since launching in summer 2025, the company manages hundreds of millions in advertising spend for major agencies worldwide. Backed by Golden Ventures and Aperiam Ventures, with a founding team that spent 15+ years in Amazon Ads and ecommerce (prior $100M+ exit).
You will join the engineering team in Toronto as an AI Agent Engineer, owning end-to-end development of agents and workflows that power Gigi's media manager. This is fundamentally a software engineering role focused on how LLMs reason, use tools, retain context, and take reliable actions on behalf of customers.
Key responsibilities:
- Ship agents end-to-end, from data and tools through reasoning and execution to customer-facing interfaces. Own capabilities like turning briefs into campaigns, investigating performance, and executing approved changes.
- Build with AI as a core part of your development workflow, using Codex, Claude Code, Cursor, and similar tools for planning, implementation, testing, debugging, and review—while remaining accountable for architecture, quality, and correctness.
- Build agent judgment: decide what Gigi should do autonomously versus escalate to humans; score actions on risk and confidence.
- Draw the line between probabilistic (LLM reasoning) and deterministic (structured software execution) work on a daily basis.
- Build reliable agentic systems with tools, retrieval, memory, orchestration, guardrails, and human approval flows that handle incomplete context, tool failures, and pausable workflows.
- Evaluate and improve agent behavior using real customer workflows, production failures, traces, and regression tests.
- Work directly with customers on calls and feedback, turning ambiguous problems into shipped product without waiting for perfect specs.
Tech stack: Python, FastAPI, LangGraph, LangChain; TypeScript, Vue, Nuxt; Java, Spring Boot; PostgreSQL, pgvector, Redis, NATS; AWS.
Success milestones: 30 days—code in production, understanding media buyer workflows; 6 months—own an agent capability outright with evaluations and production behavior; 12 months—ship something the team couldn't build without you, expand customer confidence in Gigi, and iterate on abstractions.
The company emphasizes intensity, ambiguity, and high standards. Cash compensation is competitive for stage but may not match late-stage offers; upside is meaningful ownership and equity. Engineers work together downtown Toronto but can choose remote work on Wednesdays. Benefits include catered lunch, comprehensive health/dental, and top-of-the-line Apple equipment.
REQUIREMENTS:
- Track record of building and shipping production software used by real customers; depth of ownership matters more than years.
- Strong software engineering experience with depth in backend or full-stack development (system design, APIs, data models, testing, debugging, production operations).
- Hands-on experience building with LLMs or agents—at work, open source, or substantial personal projects; ability to explain how they work, where they fail, and how to improve them.
- Comfort across the stack with real depth somewhere; no allergy to weaker areas.
- Experience using or strong desire to master AI coding tools (Codex, Claude Code, Cursor, or similar); excitement to use them daily with technical strength to review, reject weak abstractions, catch edge cases, and remain accountable.
- Curiosity about how LLMs change software products and development practices; good judgment about where AI belongs versus conventional software.
- Ability to make reasoned calls, ship, and iterate fast when wrong.
- Speed and ownership mindset; comfort using AI coding agents to move quickly while maintaining high standards for architecture, correctness, and maintainability.
NOT REQUIRED:
- AI research background or foundation model training experience.
- Prior experience with every agent framework or concept (LangGraph, orchestration, evals, retrieval, tool use learned on the job).
- Advertising or Amazon DSP domain knowledge.
BONUS:
- Production LLM work: evals, guardrails, tool use, retrieval, agent orchestration, cost/latency optimization.
- Built evaluation datasets or used production traces and customer feedback to improve agent behavior.
- Advertising or ecommerce background.
- Real-time data pipelines at scale or early-stage startup experience.
- Production systems experience where correctness, permissions, auditability, or customer trust were critical.
