Senior ML Engineer: AI Agents & Production (Remote)

Toloka; Netherlands; On-site; EUR 90,000 - 140,000; Full time

Benefits offered by this job

Remote or hybrid model

Competitive compensation including ESO

Paid PTO and benefits

Job summary

Toloka AI is seeking a Senior ML Engineer to design, train, and deploy AI agents that power Toloka's core products. You will focus on end-to-end ML tasks—including fine-tuning and RL—to build resilient agentic workflows in Python.

You will own the full lifecycle of models, from research and benchmarking to production scaling and monitoring, with a remote-first, globally distributed team and opportunities for a hybrid setup depending on location.

Qualifications

• 3+ years of experience in ML, including model training, fine-tuning and RL.
• 1+ year in building autonomous AI agents.
• Experience with frameworks such as LangChain, LlamaIndex or AutoGen.
• Practical knowledge of open-source LLMs and model distillation.
• Strong Python programming and software engineering discipline.

Responsibilities

• Train, fine-tune, and distill ML models including RL approaches for autonomous agents.
• Build and operate agentic workflows in Python with complex reasoning and human-expert interactions.
• Own evaluation, benchmarking, and cost modeling for models deployed in production.
• Manage full ML lifecycle from design to production monitoring.
• Implement observability metrics for agent logic, model performance and reliability.

Skills

ML engineering

Agent development

Python

Agent frameworks

Open-source LLMs

English fluency

Tools

LangChain

LlamaIndex

AutoGen

Job description

Toloka AI is seeking a Senior ML Engineer to design, train, and deploy AI agents that power Toloka's core products. You will focus on end-to-end ML tasks—including fine-tuning and RL—to build resilient agentic workflows in Python.

You will own the full lifecycle of models, from research and benchmarking to production scaling and monitoring, with a remote-first, globally distributed team and opportunities for a hybrid setup depending on location.