Senior AI Engineer

About us: We are a boutique software house, and by boutique, we mean that we emphasize quality over quantity. Our specialists mainly come from the JVM world, and in particular, we have taken a liking to Kotlin. Working with us, you will have a feedback session every 4 weeks and the opportunity for technical consultations directed at your current work or personal development. RadBrackets will always be a remote-first company. Client: A remote project for an international company in the fintech market. The project's goal is to build an operational infrastructure for a new generation of private markets, which combines AI-based automation with advanced regulatory knowledge to replace fragmented spreadsheets and manual processes with a single, integrated platform covering the entire fund lifecycle. In this project, TypeScript (Node.js) is the main backend technology, but your daily and most important task will be working with AI/LLM technologies, rather than standard application code development. We are looking for an AI engineer who moves proficiently within the TS ecosystem, not a classic web developer. The project is initially for 6 months with an option to extend. Work is strictly remote in English, mainly on the development of new portal functionalities. About You:

• you like challenges and new environments
• you can talk to the business, provide advice, and support software development not just through coding
• you can improve the software production process
• you support less experienced coders
• you are decisive - empathy and communication are close to you
• you communicate fluently in English

Daily tasks

• Designing and implementing AI agents and RAG pipelines – creating advanced, autonomous solutions based on LLM models using the Mastra framework.
• Model evaluation and optimization – building test sets, measuring response quality (LLM-as-judge), and implementing regression tests for prompts.
• Maintaining LLMOps processes – monitoring costs and latency of production queries, versioning prompts/models, and ensuring full observability (LLM call tracing).
• Developing semantic search – designing vector database architecture, optimizing data chunking strategies, and embedding generation.
• Building application backend in TypeScript – embedding AI logic into secure and scalable backend services (Node.js) for an international fintech platform.
• Prototyping and business collaboration – creating functional demo versions ready for presentation to the client and translating complex AI technical issues into language understandable to non-technical people.

Requirements

• You have 6+ years of experience as a TypeScript developer
• You have 2+ years of experience in building AI agents with TypeScript
• You have experience in evaluating agents and RAG pipelines (test sets, response quality metrics, LLM-as-judge, prompt regression tests)
• You understand the basics of MLOps/LLMOps: monitoring costs and latency, LLM call tracing, prompt and model versioning, production observability
• You have documented experience in semantic search (text chunking / creating embeddings)
• You possess the ability to prepare a demo ready for a sales presentation
• You have knowledge of the Mastra framework
• You can clearly communicate and explain technical gaps to non-technical people
• You have a good knowledge of system design principles, scalability, and performance optimization.
• You know how to work in CI/CD and understand the challenges associated with it.
• You possess the ability to prepare good technical documentation.
• You can design complex IT systems.

Must have: TypeScript, Node.js, SQL, Kubernetes, AWS, MySQL, CI/CD, GitHub, AI

Nice to have: React, Nest.js, Python