The Strategic Forecasting Lab is an innovation accelerator under Global Finance and Business Management / Financial Analysis focused on advancing JPMC’s forecasting capabilities. The lab’s work has historically revolved around the design and implementation of F3, the Firmwide Forecasting Framework – JPMC’s federated, end-to-end, multi-purpose forecasting platform, and in 2017, the Lab is shifting priorities to incorporating advanced analytical tooling into F3.
Machine Learning Associate
The candidate will work in the Research arm of the Lab, developing innovative quantitative solutions across all lines of work of the lab. The candidate must have strong technical skills in statistical modelling and scientific computing to contribute to the development of core components of the platform, as well as excellent communication skills to present results. The candidate will largely act as a domain expert, but hands-on involvement in prototyping different solutions is expected. Responsibilities:
- Research quantitative methods and approaches to solving problems in the lines of work of the lab, prototyping solutions, and partnering with development teams and technology for their implementation.
- Deeply understand current forecasting models and the way they are used for analysis, identifying opportunities for improvement of the existing process– with particular emphasis in the introduction of new quantitative methods rooted in machine learning.
- Document innovations for intellectual property filing and model review, where applicable. Prepare plans, status, and findings for management and regulators.
- Participate in hands-on review of tools, documentation, process, and data to assess business requirements and conceptual and financial soundness of the implementations, as well as compliance with rules and best practices.
- Stay current with emerging developments in models, approaches and requirements, both regulatory and others, and oversee the corresponding changes in models, systems, and processes.
- Partner with internal and/or external Project Management Office resources to enhance end-to-end forecasting process and achieve continuous improvement.
- A balanced blend of pragmatism and solid theoretical grounds (yes, we do believe Ali Rahimi is mostly right, but at the same time, also understand that mankind have been building ships before hydrodynamics had been fully understood). Empirical rigor, and a deeply internalized sense of the scientific method, having zero tolerance for bad science or improperly substantiated claims
- An inquisitive mind with a healthy habit to dig deep in the questions to understand the context, ask hard questions, and reframe problems when needed
- An advanced degree (preferably PhD or equivalent level of knowledge) in a quantitative field (e.g. computer science, finance, economics, electrical engineering, statistics, etc.)
- Excellent programming skills, preferably with experience in advanced R, Python
- Knowledge of Machine Learning and some typical stacks for machine learning (e.g. scalable implementations of boosting, Tensorflow, MXNet, MLlib, Scikit-learn, tooling in R's MachineLearning task view), and/or model computational statistics and typical frameworks (e.g. implementations of GAM's and GAMLSS methods in R, Stan). Strong understanding of the different methods – at the very minimum at the level of “intelligent user” of packaged methods, but ideally also knowing the algorithmic details at the level of being able to modify existing learning methods and even write new ones from scratch
- Knowledge of typical big data tooling, preferably with experience in Spark, Impala. Strong knowledge of data engineering – including databases (both relational and not relational), data I/O, cleansing, and transformation is highly desired, including advanced SQL analytical extensions for longitudinal data
- Knowledge of Econometrics (Time Series Analysis, Longitudinal Data Analysis) desirable
- Polished communications and Data Visualization highly desirable. Familiarity with patterns and typical libraries (e.g. D3.js, Material) and application frameworks (e.g. R’s Shiny, Angular, Polymer etc.) commonly used to build highly interactive analytical applications
- Strong analytical and problem solving, and more importantly, problem finding abilities
About J.P. Morgan Chase & Co:
J.P. Morgan serves one of the largest client franchises in the world. Our clients include corporations, institutional investors, hedge funds, governments and affluent individuals in more than 100 countries. J.P. Morgan is part of JPMorgan Chase & Co. (NYSE: JPM), a leading global financial services firm with assets of $2.1 trillion. The firm is a leader in investment banking, financial services for consumers, small business and commercial banking, financial transaction processing, asset management, and private equity. A component of the Dow Jones Industrial Average, JPMorgan Chase serves millions of clients and consumers under its JPMorgan and Chase, and WaMu brands. J.P. Morgan offers an exceptional benefits program and a highly competitive compensation package. J.P. Morgan is an Equal Opportunity Employer