Job Description

Project Details

Natural Language processing v2.0

The bank generates a vast amount of data like customer data, logs from their financial products, transaction data that can be used in order to support decision making, together with unstructured data, like social media data and customer emails etc. As the unstructured data and the technology to process it grows, there is a natural demand to tap the insights from this data.

The Bank wants to classify customer emails into different issue buckets to act on them quickly. In this project the candidate has to new age DL algorithms to train supervised and unsupervised algorithms to achieve ~90% accuracy.

Key Deliverables

Regular discussion with Banks AI team to finalize the approach and regular communication on the project track report
EDA & technique finalization
Build the model custom trained model on vocab (supervised and unsupervised)
Validate the model by choosing the right metrics
Responsible for the deployment pipeline on cloud
Experience and Profile

4-8 years of relevant industry experience (NLP & Deep learning), with excellent mathematical, programming and problem-solving skills, as well as a passion for AI and research
The candidate should be an expert at analysing large amounts of raw information to find patterns and extract valuable insights

For successful execution of the job, the candidate should possess the following:

Knowledge /Technical Capability :
Fluency in Python – hands on relevant experience
Experience with NLP libraries like core NLP, spacy, genism, etc
High-performance processing with robust analytic data pipeline
Fluency in at least one deep learning framework: PyTorch, TensorFlow / Keras
Advanced knowledge of the HuggingFace libraries (transformers and tokenizers)
Strong analytical & problem solving mindset with an attitude of “professional scepticism”
Good communication (both verbal & written)
Ability to manage complex client situations
Ability to multitask
Ability to handle pressure and meet deadlines