Description

We are the AI team within Sber's Self-Service Device Network Management division (ATMs and other devices), responsible for implementing machine learning and GenAI solutions into mass operational processes for customer self-service. We develop AI solutions for production using big data. Within the team, we build full-cycle ML and GenAI systems: from processing data streams and instructions to AI agents that assist employees and automate decision-making. Currently, we are strengthening the team and looking for a Tech Lead MLE who will take responsibility for engineering quality in classical ML and GenAI areas: RAG systems, AI agents, integration of LLMs into critical business processes, and deployment of solutions to PROD.

Responsibilities

Technical Leadership and Architecture

• Designing end-to-end ML / GenAI systems: from business task to PROD

• Selecting solution architecture (LLM, RAG, agents, classical ML)

• Making key technical decisions and being accountable for them

• Code review, establishing development standards within the team • Mentoring MLE / DS / analysts

Development of GenAI Solutions

Development and implementation of AI agents for internal and customer scenarios

• Designing RAG systems (including complex graphs and multi-agent schemes)

• Integrating LLMs into the business processes of the self-service device management division (ATMs and others)

• Fine-tuning models for the domain (SFT / LoRA / embeddings / rerankers)

• Optimizing latency, cost, and stability of LLM solutions

ML and Analytics

• Developing and maintaining ML models (classical ML + DL)

• Feature engineering, validation, A/B testing

• Monitoring model quality, evaluating and defending AI impact

• Working with time series, anomalies, forecasting

Deployment and Operations

• Developing AI solutions for PROD

• Building ML pipelines (training, retraining, monitoring)

• Integration with existing IT architecture

• Ensuring reproducibility and fault tolerance of solutions

Requirements

• Technical leadership: code review, architectural decisions, mentoring

Knowledge and Skills

• Python (production-level, PEP8, tests, architecture)

• SQL (complex queries, window functions)

• Classical ML (GBM, DL, Time Series — at a confident production level)

• LLM / GenAI: • Prompt engineering

• RAG (embeddings, rerankers, chunking, retrieval strategies)

• AI agents (LangChain / LangGraph / ReAct / Tools)

• Handling hallucinations, latency, cost control · MLOps:

• Docker, CI/CD for ML •

REST / Async API (FastAPI)

• Apache Spark (Pyspark, Spark SQL) – development and optimization of batch pipelines

• Hive/HDFS/data marts: preparation and maintenance of data marts for ML models and GenAI solutions

Conditions

• participation in a unique and large-scale AI development project
• comfortable modern office near Kutuzovskaya metro station
• work format: hybrid
• annual salary review and yearly bonus
• corporate gym and relaxation areas
• more than 400 educational programs from SberUniversity for professional and career development
• extended health insurance (DMS), preferential insurance for family, and corporate pension program
• flexible mortgage discount equal to 1/3 of the Central Bank's key rate
• free SberPrime+ subscription, discounts on products from partner companies
• referral bonus for recommending friends to Sber's team.