Who we'd reach at ЯндексFound and reached by Power Apply
Best leadTeam lead
Runs the team day to day
Recruiter
In-house, owns the opening
Recent hire
Joined the team recently
From your own Gmail and LinkedIn.
About the role
This is a data scientist role at Yandex for people who build machine learning components, quality metrics and models for AI products, ad moderation, anti-fraud and search. The listing does not state the location, the country or the pay, and it is written in Russian. The work covers evaluating answer quality for generative models and agent scenarios, moderation and fraud detection, and building benchmarks for generative model research.
What you’ll do
- Evaluate the quality of answers and agent scenarios for generative models, and prepare data for model improvement, including ethics and safety tasks.
- Develop moderation and anti-fraud systems for advertising, automate processes and detect abuse.
- Find causes of content playback and delivery problems using logs and telemetry, and build ML components that assess and improve quality.
- Build metrics that drive key product decisions, and train ML models in some teams.
- Create benchmarks for generative models that test reasoning, factual knowledge, ethics, cultural context and responses to provocation.
What they’re looking for
- Python and SQL or YQL, with experience handling large data volumes.
- Probability theory, mathematical statistics and hypothesis testing.
- Understanding of ML principles and the ability to match solution complexity to the problem.
- Ability to turn product and business questions into formal problems, choose quality metrics and analyse errors.
- Writes maintainable code and can organise data processing.
- Works independently, explains conclusions clearly and collaborates with developers and stakeholders.
- Has trained and deployed models, or worked with LLMs, NLP, computer vision or AI quality evaluation systems.
Skills
Questions about this role
What does the hiring process involve?
A discussion of your experience and interests, technical sessions, and meetings with suitable teams, followed by a discussion of the job offer once the stages are passed.
Which programming languages and query languages are required?
Python, and SQL or YQL.