Description
IT B2C is the largest ecosystem within Sber. We are over 8000 people across 18 cities in Russia. We are engaged in the development and enhancement of retail solutions, helping to make the Bank's services more accessible, secure, and convenient.
We are looking for you!
For the development of technologies based on GigaChat and to assist our retail clients with financial solutions, we are launching a new project and assembling a team for it. We are looking for a developer to join this team for project-based work. The primary programming language is Python.
The project's goal is to create an AI consultant for retail lending products.
Responsibilities
- Communication with colleagues and clients, clarifying business requirements, refining specifications.
- Studying the business processes of retail lending products.
- Searching for information in documentation (as much as necessary to solve our tasks).
- Participation in code reviews.
- Application development, including typical tasks:
- Using prompt engineering methods to improve product response quality.
- Implementing integration tasks (calling API services to obtain client information, reading from and writing to dialogue databases, vector storage).
- Transforming API responses into text for passing to the LLM.
- Using LLM to gather necessary information for the process from client dialogue.
- Connecting platform services.
Requirements
- Python knowledge at a middle+ level.
- Understanding of the architecture and principles of large language models.
- Knowledge of algorithms and data structures.
- Experience in developing web services on FastAPI/Django/Flask.
- Experience with Docker/Kubernetes.
- Prompt engineering experience.
- Experience in finding errors, debugging applications.
- Experience in code profiling and identifying performance bottlenecks.
- Experience participating in code reviews.
- Git skills.
Will be a plus:
- Knowledge of the langchain/gigachain framework (Chains, LCEL, Retrieval, Agents).
- Experience with Platform V (EFS)
- Understanding of LLM model evaluation methods (Human evaluation, Perplexity, BLEU, ROUGE, others).
- Knowledge of NLP basics, processing modules (nltk).
- Experience conducting load testing and evaluating results.
- Experience writing and refactoring code using an IDE (IDEA, PyCharm).
- Experience using project management systems (Jira).
- Willingness to share experience with colleagues.
Conditions
- Office work format
- Annual bonus and yearly salary review
- Extended VHI from day one and preferential insurance for family
- Sber Corporate University, internal educational platform, participation in IT conferences
- Office on Kutuzovsky Prospekt with rest zones and a gym
- Preferential mortgage at Sber, corporate pension program, SberPrime+ subscription, discounts from partners and services of the group of companies
