AVP/Senior Associate, Data Scientist, Technology and Operations
Group Technology and Operations (T&O) enables and empowers the bank with an efficient, nimble and resilient infrastructure through a strategic focus on productivity, quality & control, technology, people capability and innovation. In Group T&O, we manage the majority of the Bank's operational processes and inspire to delight our business partners through our multiple banking delivery channels. Our goal is to build T&M NLP core engine to handle all T&M related ML use cases, e.g, sentiment analysis, text summarization, classification, NER, trade info extraction, etc. We have extended BERT base vocabulary with T&M domain specific words, and completed T&M BERT pre-training with training dataset of 5,000,000 plus T&M documents. Our next milestone is to build a generic T&M NER engine to extract trades details from free-text documents, e.g, emails, chats, pdf, etc. The traditional NER engine, i.e, Spacy, could only recognize the default entities, but not able to understand the trades entities, e.g, forex rate, settlement date, coupon rate, counterparty, etc. Responsibilities
- Work with large and complex financial datasets to build end-to-end data science solutions.
- Deploy machine learning products into production and perform sub-system integration as required.
- Refactor and document code into reusable libraries, APIs, and tools.
- Automate CI/CD pipeline for model training, testing and deployment.
- Assist organizational adoption of industry leading ML ecosystem (feature/eval store, model zoo)
- Optimize machine learning algorithm efficiency (GPU distributed computing, concurrent programming)
- Experience building scalable machine learning system architectures (microservice, distributed) and big-data pipelines in production.
- Good understanding of the data science production life cycle with demonstrable experience working with structured, semi-structured and unstructured data.
- Excellent software skills (Python, SQL, bash) and knowledge in design patterns and code optimization.
- Good grasp of Machine Learning models and concepts, their mathematical underpinnings, and trade-offs (model selection, tuning, problem formulation, drift, semi-supervised learning)
- Experience using NLP techniques (NER, Sentiment Analysis, Topic Modelling, transformers)
- Experience using machine learning frameworks (TensorFlow, Pytorch, Spark, Hydra)
- Familiar with software development best practices and tools (Agile, TDD, Jira, Jenkins, Git)
- Familiar with Linux operation system
- Financial domain expertise (Treasury & Markets)
- Experience with statistical modelling and time series analysis.
We offer a competitive salary and benefits package and the professional advantages of a dynamic environment that supports your development and recognises your achievements.