Description
The "News Monitoring" service team is looking for a Middle ML Engineer.
We are engaged in developing NLP models for analyzing news streams. Key tasks include: NER, entity linking, classification. We are opening a new OCR direction. It will be necessary to automate the incoming flow of documents in various formats.
Responsibilities
Tasks:
- We have an end-to-end real-time news processing pipeline - from parsing to delivery to users. A lot of interesting things happen inside, but it's not always optimal. It is expected that a competent MLE will help us with this.
- There are tasks related to organizing data storage (currently there is a slow storage in PostgreSQL, fast storage in ElasticSearch), microservice architecture (there are many Docker containers, we can try to move everything to Kubernetes), and model deployment (we want to set up automatic rollout to torchserve via gitlab-ci + ansible)
Requirements
Requirements:
- Confident Python, SOLID, ability to understand others' code
- Databases: PostgreSQL, knowledge of ElasticSearch is a plus
- ML/DL - sklearn/PyTorch, knowledge of transformers is a plus
- Gitlab CI, Docker, ansible
- Preferably have experience with wrapping PyTorch models in onnx/tensorrt
Conditions
Conditions:
- ML Space for model training
- Largest DS&AI community - over 600 DS in the bank, including:
- Regular exchange of knowledge, experience, and best practices
- Ability to reuse code and libraries from a centralized repository
- Interactive lectures and master classes from leading universities and experts from technology companies
- Digest of the latest developments in the field of DS&AI
- Regular internal meetups
