AI Production Engineer

MAKE YOUR STORY REAL

At Siili, we don't just build AI – we run it. We've seen firsthand that capabilities needed to run a compelling proof of concept are fundamentally different from those required to operate AI reliably at scale. That's why we're building our Managed AI practice: to bridge the gap between "it works in the PoC" and "it works in our business."

As an AI Production Engineer, your work focuses on operating AI systems in production — not building experimental models or prototypes. You will ensure that existing AI applications stay reliable, accurate and cost-efficient over time. You'll work as part of our Managed Services business line, alongside developers, data, automation and cloud experts. Your focus will be on production AI: building AI systems, monitoring model performance, detecting drift, optimizing inference, and responding when things break.

What you'll do:

• Monitor and maintain production AI systems, ensuring reliability, performance, and quality under SLAs.

• Implement end-to-end AI solutions to production deployment

• Implement model monitoring solutions: drift detection, performance tracking, and automated alerting.

• Optimize LLM applications for production: latency, throughput, token usage, and cost efficiency.

• Build and maintain evaluation frameworks to catch quality degradation before users do.

• Troubleshoot AI-specific incidents: hallucinations, embedding quality issues, retrieval failures, and unexpected model behavior.

• Implement retraining pipelines and manage model versioning in production environments.

• Optimize RAG systems: chunking strategies, embedding refresh, vector database performance, and retrieval quality.

• Collaborate with client teams to improve their AI systems based on production insights.

• Document operational runbooks and contribute to Managed AI best practices.

• Drive continuous operations improvements by identifying recurring issues, automating routine tasks, and ensuring long-term system reliability.

What we're looking for:

We are looking for an AI Production Engineer who understands what it takes to run AI in production – not just build it.

We hope you recognize yourself in most of these:

• 2+ years of experience in AI/ML engineering, with hands-on production deployment and operations experience.

• Strong understanding of ML concepts and the ways models fail in production: drift, degradation, edge cases.

• Experience with LLM applications in production: prompt management, evaluation, monitoring, and optimization.

• Knowledge of MLOps practices: CI/CD for ML, model versioning, A/B testing, and staged rollouts.

• Familiarity with RAG architectures and their operational challenges: embeddings, vector databases, retrieval tuning.

• Hands-on experience with monitoring and observability tools.

• Ability to diagnose and resolve production issues under pressure.

• Good communication skills – able to explain AI behavior to non-technical stakeholders and write clear incident reports.

• Curiosity about why models behave unexpectedly and drive to prevent it from happening again.

• Fluent in Finnish and English.

Relevant technologies:

• LLMs and LLM middleware: GPT, Claude, open-source models, LiteLLM, vLLM

• Agent technologies: MCP, agent/workflow frameworks (LangGraph, LlamaIndex, Haystack, Magentic, Strands)

• Platforms, e.g. Azure AI Foundry, Amazon Bedrock, Snowflake and Databricks

• Cloud: Azure, AWS, GCP

• Python

• Vector databases, e.g. Pinecone, pgvector

• Monitoring: Prometheus, Grafana, Datadog, LLM-specific observability (Langfuse, Phoenix)

• Containers and orchestration basics

What we offer:

• A technically strong and collaborative community: You'll work alongside experienced data professionals who support each other, share insights, and aim for excellence in everything we do. Our culture values respect, autonomy, and peer-driven improvement.

• Deep AI production expertise: You'll develop specialized skills in AI operations that most engineers never get – because most companies don't have production AI systems yet. You’ll be shaping how AI is run in the real world.

• Real-world impact, not just buzzwords: You'll help deliver and run scalable, meaningful Data & AI solutions with clear ties to business strategy – not just experimental prototypes.