Description
We are the team of GigaChat Data ML Engineers.
Our team prepares data for training the entire GigaChat LLM model lineup (GigaChat, GigaChat Vision, GigaChat Audio, Giga Embeddings, etc.). We collect data both for base models and for their adaptation to specific scenarios:
- Core models – base models
- B2C products – WEB version of GigaChat, personalized intelligent assistants
- internal banking tasks – automation of document processing, analysis of financial data, employee support
- external customers – custom solutions for corporate clients, integration of LLM into business processes.
Responsibilities
Tasks:
- collecting data for model alignment training: Supervised Finetuning (SFT), Reinforcement Learning (RL, DPO, PPO, GRPO)
- improving data preprocessing pipelines for annotation and training
- working closely with the ML team, LoRA-finetuning
- interaction with the assessor/annotator team (monitoring quality, adjusting instructions)
- calculating metrics and making decisions to improve data quality based on them.
Requirements
What is important to us:
- very high level Python knowledge
- experience working with proprietary and open-source LLMs (Llama, Mistral, Qwen, Google Gemini, OpenAI GPT, Anthropic, etc.)
- understanding of data workflows for large language models
- ability to solve problems and organize work in a multitasking environment.
Will be a plus:
- experience with LLM agents (LangChain, LangSmith, LangGraph, FAISS, RAG, ChromaDB, etc.).
Conditions
We offer:
- largest DS&AI community – over 600 DS specialists from the bank
- digest of the latest developments in DS&AI and reports from the world's largest conferences
- opportunity to choose a convenient work format: hybrid or office
- comfortable modern office: Kutuzovskaya metro station, Kutuzovsky Prospekt, 32
- annual salary review, annual bonus
- corporate gym and relaxation areas
- more than 400 educational programs from SberUniversity for professional and career development
- extended health insurance, preferential family insurance and corporate pension program
- mortgage more favorable by up to 7% for every employee
- free SberPrime+ subscription, discounts on products from partner companies
- referral bonus for recommending friends to the Sber team.
