machine learning engineer

Toloka creates data that powers leading GenAI models and innovations. The company combines experts, crowds, and a technology platform to train AI models and evaluate their efficacy and safety. Its ML team builds machine-learning products that make the platform faster, cheaper, and more reliable.

• Train, fine-tune, and distill ML models, including through Reinforcement Learning (RL), to power autonomous AI agents;
• Build and operate agentic workflows in Python, handling complex reasoning and hybrid human-expert interactions;
• Own evaluation and benchmarking, select foundational models, and establish cost models;
• Manage the full ML lifecycle from solution design through production deployment and monitoring of real-time signals;
• Implement observability metrics for agent logic, model performance, and system reliability.

• 3+ Years of experience in ML, with a strong background in model training, fine-tuning, and Reinforcement Learning (RL);
• 1+ Year of practical experience building, evaluating, and launching autonomous AI agents;
• Proven experience with agentic frameworks such as LangChain, LlamaIndex, or AutoGen;
• Practical knowledge of model distillation and adapting open-source models such as Llama and Mistral;
• Advanced Python proficiency and disciplined software engineering standards applied to ML;
• Ability to work across the entire chain from research to production operations;
• Fluent English at B2 level or above.

• Competitive compensation package including base salary, bonus, and ESOP;
• Paid PTO and benefits vary depending on location;
• IT setup and home office allowances.