At the AI Optimizer team, our day is usually full of R&D challenges. Have you ever encountered a situation where you need to expand your AI infrastructure so that the applications can automatically pick the right large language models (LLMs) that are both more cost-efficient and better performing? Most of us probably do nowadays, or at least understand the complexity of making such decisions while keeping track of our cloud budget.

One of the team's responsibilities is ensuring that whenever a customer makes AI-related decisions regarding their K8s infrastructure, they are implemented automatically without unnecessary costs or hassle. This is just one small piece of a bigger puzzle. To get into a more detailed perspective, ask yourself the following questions:

• How often do you use LLMs?
• What is the least expensive LLM you can pick for a given prompt without degrading the quality of the response?
• How much do your applications cost per 1 million tokens and how can you improve it?
• Which API keys have the biggest waste?
• How can you improve your frequently running prompt to use fewer tokens?
• What is fine-tuning and how to do it efficiently?
• What is a transformer?

These are just several of the many questions that are part of the daily work of this team.

Being part of this team would involve design and decision-making end-to-end while collaborating with colleagues from other teams. CAST AI, being a technical product, encourages not only coding something as written in the JIRA ticket but also coming up with new features and potential solutions to customers' problems. Given that the team is working on a technical greenfield project, you will have the opportunity to impact it in many ways positively.

• Evaluate and Analyze LLM performance
• Fine-Tune LLMs
• Optimize AI Models for Cost Efficiency
• Develop and implement data science solutions
• Monitor and improve AI systems
• Stay up to date with industry trends.

• GoLang is our main language, while Python is an accepted alternative for some cases
• ClickHouse and PostgreSQL for persistence
• GCP Pub/Sub for messaging
• gRPC for internal communication
• REST for public APIs
• Kubernetes which our product is evolving around
• AWS, GCP, and Azure cloud providers, which are currently supported in our platform
• We use GitLab CI with ArgoCD as our GitOps CD engine
• Prometheus, Grafana, Loki, and Tempo for observability.

• You have to be physically in any of the European countries GMT 0 to GMT +3
• Strong software engineering skills
• Minimum of 3 years of hands-on experience in Data Science and Machine Learning, with a proven track record, demonstrated through a robust portfolio of projects
• Strong English skills
• Strong verbal and written communication skills
• Ability to work independently or with a group.

• Team of highly skilled professionals to work with and learn from
• Impact and visibility. Our organization is flat, getting in touch with CEO or CTO is a common practice here
• Monthly salary from €6000 to €8000 (gross) depending on the level of experience
• Short feedback loop. We have an obsession with customer satisfaction. The ship features fast and gets instant feedback. Feature projects tend to be completed in 1 to 4 weeks, depending on the scope
• Flexible working hours. We deliver instead of sitting in the office 8 to 5
• Skin in the game. Every employee gets a share of the company
• Time to focus on work with a minimum overhead of meetings, bureaucracy, etc.
• 10% time to focus on self-improvement or personal projects.