Lead AI Engineer — Agentic Products & Platform

About the Role

Simulmedia is building AI into how TV advertising is planned, activated, and measured. This role is for the engineer who builds it.

You are a strong backend engineer who has gone deep on AI and agentic systems: someone as comfortable designing a production service architecture as reasoning about LLM context windows, tool use, and agent failure modes. You will own the design, implementation, and ongoing reliability of Simulmedia's agentic AI products, working across Product, Data Science and Engineering teams to bring these systems to life.

You are the technical owner of Simulmedia's AI systems: the person who drives how they are built, how they behave in production, and how the rest of the team builds around them.

This is a lead individual contributor role with broad technical influence. You are not managing a team. You are the person others look to when the question is how do we actually build this.

What You'll Own

Agentic AI Products

• Design and build AI agents embedded in the product across campaign management, workflow automation, audience intelligence, and other areas on the roadmap
• Own the full engineering stack for each agent: tool definitions, orchestration logic, context management, output validation, guardrails, and human-in-the-loop design
• Translate product and data science intent into working, observable, production-grade agentic systems
• Set the technical patterns and reference implementations that the broader engineering team builds around

AI Platform and Integration Layer

• Build and own the shared infrastructure that agentic features run on: LLM API integration, prompt management, agent memory and state, evaluation harnesses
• Define the interfaces between agentic systems and the rest of the platform: how agents consume data, trigger actions, and return results to product surfaces
• Work with Data Engineering to ensure agents have access to the right data at the right latency from the Databricks pipeline
• Work with Full Stack engineers to surface agent capabilities in product UIs, APIs, and integrations: you own the backend contract, they own the surface

Reliability and Observability for AI Systems

• Define what healthy looks like for an agentic system and instrument accordingly: response quality, latency, failure rates, fallback behavior
• Build evaluation pipelines to catch regressions before they reach users
• Own incident response for AI system failures, including participation in 24/7 on-call rotation: you understand the system well enough to diagnose and fix quickly

Cross-functional Technical Collaboration

• Be the go-to technical resource for Data Scientists who want to embed their models and outputs into product workflows: not owning the models, but owning the engineering that makes them usable in production
• Help Full Stack and Data Engineering teams build AI-aware features confidently, by establishing clear patterns and being available to unblock them
• Document decisions, patterns, and system designs so the team compounds on your work rather than depending on you to re-explain it

What We're Looking For

Background

• 6+ years in backend or platform engineering, with meaningful experience building AI or agentic systems in production
• A track record of shipping things: not just prototypes, but systems that work reliably, that other engineers maintain, and that users depend on
• Experience working at the intersection of engineering and data science: you've taken research-stage work and turned it into production systems

AI and Agentic Systems

• Deep hands-on experience with LLMs in production: tool use, function calling, RAG, multi-step reasoning, prompt engineering, and the failure modes that come with each
• Experience building agentic workflows, whether with frameworks like LangGraph, CrewAI, AWS Bedrock or custom orchestration, and strong opinions about the tradeoffs
• Understanding of how to design human-in-the-loop systems: when to intervene, how to make intervention low-friction, and how to build user trust over time
• Familiarity with evaluation and testing approaches for non-deterministic systems

Backend Engineering

• Strong fundamentals: API design, service reliability, async patterns, observability, and writing code that other engineers can understand and extend
• Comfortable with Python and TypeScript as primary languages; Go experience is a plus
• Enough data engineering fluency to work confidently with pipelines, SQL, and a warehouse environment (we use Redshift and Databricks)

Communication

• Business-level written and spoken English is required. This role involves regular collaboration with Ukraine-based and US-based stakeholders across Product, Data Science, and Commercial teams.
• Clear and precise technical communication: you can explain what an agent is doing, why it failed, and what you're doing about it to both engineers and non-engineers

Technology Context

Our AI stack is being shaped in part by whoever takes this role. The broader platform stack: Python, Ruby, Go, TypeScript/Next.js, Databricks, PostgreSQL, AWS, Kubernetes, GitHub Actions.

Ways of Working

• You build: not just direction-setting, not just prototyping. You own working production systems.
• Collaborative by design: the role only works if Data Scientists, Full Stack engineers, and Data Engineers can build with confidence around what you have created
• Prototype-first: working code over slide decks, always
• Documentation as a multiplier: your patterns and decisions should compound across the team, not live only in your head