Tech Lead / Senior ML Developer for the Iron Intern Team

Tech Lead / Senior ML Developer for the Iron Intern Team

We are creating the Iron Intern – an AI agent that will accelerate development across all Yandex projects and help developers solve various routine tasks faster. We are looking for an NLP developer with experience in building NLP projects, ready to develop complex SOTA systems for writing code, gathering requirements, and searching for expert knowledge, make unconventional decisions, and interested in influencing development.

We offer you: * Creating an AI developer. You will gradually take it on the whole journey from intern to senior, while dramatically accelerating the company's core processes along the way * Complex architectural and technological challenges * Extensive knowledge and experience in developing AI products from the creators of YandexGPT, Neuro, Yandex Browser neurotechnologies, and the voice assistant Alisa

What tasks await you

Teaching LLM agents to write code independently You will research SOTA approaches to developing NLP agents, systematize articles and create strategies, and based on the best one, develop the Iron Intern.

Building an environment where agents can test their changes Like a human developer, it's difficult for an LLM agent to write correct code immediately. It needs an interface for debugging solutions. For this task, you need to develop a testing ground where the agent can test its changes and commit them.

Developing agents to gather knowledge about code from developers and the intranet To make correct decisions, a modern developer needs more than just code. They need expert knowledge that exists on the internet, the intranet, and in people's heads. An important task is to create an agent that will collect all the necessary information to solve a task. A good strategy is to ask your non-iron colleague (but you shouldn't ask the same question too often).

More about ML at Yandex – on the channel Yandex for ML

We expect that you

• Have successfully developed and implemented NLP products
• Understand the fundamentals of modern NLP architectures
• Are ready to participate in the entire AI agent development pipeline: from design and requirements gathering to model fine-tuning and implementation