JPMorganChase is looking to solve real-world problems using machine learning methods, including natural language processing, speech analytics, and recommendation systems, to serve consumers, small businesses, municipalities, and non-profits.
Requirements
- Good understanding of the latest advancement of NLP concepts, such as the transformer architecture and knowledge distillation
- Experience in classical ML techniques including classification, clustering, optimization, cross validation, data wrangling, feature selection, and feature extraction
- Ability to design experiments — establish strong baselines, choose meaningful metrics, and evaluate model performance rigorously
- 2 years of hands-on experience with virtual assistant model development and optimization
- Familiarity with continuous integration models and unit test development
- Experience with A/B experimentation and data/metric-promoten product development
Responsibilities
- Develop state-of-the art machine learning models to solve real-world problems and apply it to tasks such as NLP, speech recognition and analytics, or recommendation systems
- Choosing, extending and innovating ML strategies for various banking problems
- Analyzing and evaluating the ongoing performance of developed models
- Collaborate with multiple partner teams such as Business, Technology, Product Management, Legal, Compliance, Strategy and Business Management to deploy solutions into production
- Learning about and understanding our supported businesses in order to promote practical and successful solutions
Other
- BS with 5+ years, or MS with 3+ years of hand-on industry experience in Machine Learning - Deep Learning
- Solid written and spoken communication skills
- Scientific thinking with the ability to invent and to work both independently and in highly collaborative team environments
- Comprehensive health care coverage, on-site health and wellness centers, a retirement savings plan, backup childcare, tuition reimbursement, mental health support, financial coaching and more
