Applied Ai/ml & Causal Inference - Senior Associate

JPMorgan Chase · Banking · Jersey City, NJ +1 · Asset & Wealth Management

Senior Associate role focused on building and deploying causal and predictive ML models for financial services clients, with a strong emphasis on causal inference, LLMs, and agentic AI systems. The role involves full lifecycle ownership from problem framing to production, including model quality, MLOps, and applied research translation.

What you'd actually do

Frame ambiguous client and operational questions as causal problems — distinguishing prediction from intervention, identifying confounders, and designing the right estimand with Private Bank business leads.
Design, build, and deploy end-to-end ML and causal inference solutions: uplift and heterogeneous treatment effect models, observational causal studies (DiD, IV, RDD, synthetic controls, doubly robust estimation), experimentation, and classical/generative ML where appropriate.
Own model quality, identification assumptions, sensitivity analysis, evaluation frameworks, monitoring, and post-deployment iteration.
Drive productionization and MLOps practices in collaboration with engineering across distributed data infrastructure.
Track applied research in causal ML, double machine learning, and agentic/LLM systems; translate promising work into production-ready solutions.

Skills

Required

Master's or PhD in Computer Science, Statistics, Economics, Applied Math, Data Science, or a related quantitative field.
3+ years of hands on Machine Learning experience in production environments, with a substantial portion focused on causal inference.
Deep expertise in causal inference methods: potential outcomes framework, propensity score methods, instrumental variables, difference-in-differences, regression discontinuity, synthetic controls, doubly robust and double/debiased ML estimators, and uplift / heterogeneous treatment effect modeling.
Demonstrated experience designing and analyzing experiments (A/B tests, switchback, quasi-experiments) and reasoning carefully from observational data when experimentation is infeasible.
Hands-on experience with LLMs and agentic AI — fine-tuning, RAG pipelines, prompt engineering, and the design and deployment of multi-step / tool-using agents in production.
Strong Python skills; proficiency with causal libraries (DoWhy, EconML, CausalML) alongside PyTorch, scikit-learn, and modern LLM/agent frameworks.
Experience with large-scale data processing: Spark, Hive, SQL.
Proven ability to communicate causal assumptions, limitations, and findings to non-technical stakeholders.

Nice to have

Financial services experience — wealth management, lending, or advisory.
Bayesian and hierarchical modeling; structural causal models; sequential decision-making / contextual bandits.
Experience applying causal reasoning to LLM and agent evaluation — counterfactual eval, off-policy estimation, or treatment-effect framing of agent interventions.

What the JD emphasized

substantial portion focused on causal inference
Deep expertise in causal inference methods
Hands-on experience with LLMs and agentic AI

Other signals

Causal inference models
Production deployment at scale
LLM and agentic AI

Read full job description

As a Senior Applied AI/ML Associate within the Global Private Bank, you will own the full lifecycle of high-impact causal and predictive models serving clients across wealth management, deposit, lending, and advisory — from problem framing with business stakeholders through production deployment at scale. You will tackle some of the most data-rich, complex client problems in financial services, where rigorous causal reasoning — not just predictive accuracy — drives the decisions that matter.

Job Responsibilities

Frame ambiguous client and operational questions as causal problems — distinguishing prediction from intervention, identifying confounders, and designing the right estimand with Private Bank business leads.
Design, build, and deploy end-to-end ML and causal inference solutions: uplift and heterogeneous treatment effect models, observational causal studies (DiD, IV, RDD, synthetic controls, doubly robust estimation), experimentation, and classical/generative ML where appropriate.
Own model quality, identification assumptions, sensitivity analysis, evaluation frameworks, monitoring, and post-deployment iteration.
Drive productionization and MLOps practices in collaboration with engineering across distributed data infrastructure.
Track applied research in causal ML, double machine learning, and agentic/LLM systems; translate promising work into production-ready solutions.
Partner with the broader JPMorganChase AI/ML community, model risk, compliance, and peer LOBs to align on standards and amplify firm-wide impact.

Required Qualifications, Capabilities, and Skills

Master's or PhD in Computer Science, Statistics, Economics, Applied Math, Data Science, or a related quantitative field.
3+ years of hands on Machine Learning experience in production environments, with a substantial portion focused on causal inference.
Deep expertise in causal inference methods: potential outcomes framework, propensity score methods, instrumental variables, difference-in-differences, regression discontinuity, synthetic controls, doubly robust and double/debiased ML estimators, and uplift / heterogeneous treatment effect modeling.
Demonstrated experience designing and analyzing experiments (A/B tests, switchback, quasi-experiments) and reasoning carefully from observational data when experimentation is infeasible.
Hands-on experience with LLMs and agentic AI — fine-tuning, RAG pipelines, prompt engineering, and the design and deployment of multi-step / tool-using agents in production.
Strong Python skills; proficiency with causal libraries (DoWhy, EconML, CausalML) alongside PyTorch, scikit-learn, and modern LLM/agent frameworks.
Experience with large-scale data processing: Spark, Hive, SQL.
Proven ability to communicate causal assumptions, limitations, and findings to non-technical stakeholders.

Preferred Qualifications, Capabilities, and Skills

Financial services experience — wealth management, lending, or advisory.
Bayesian and hierarchical modeling; structural causal models; sequential decision-making / contextual bandits.
Experience applying causal reasoning to LLM and agent evaluation — counterfactual eval, off-policy estimation, or treatment-effect framing of agent interventions.

Job Responsibilities

Frame ambiguous client and operational questions as causal problems — distinguishing prediction from intervention, identifying confounders, and designing the right estimand with Private Bank business leads.
Design, build, and deploy end-to-end ML and causal inference solutions: uplift and heterogeneous treatment effect models, observational causal studies (DiD, IV, RDD, synthetic controls, doubly robust estimation), experimentation, and classical/generative ML where appropriate.
Own model quality, identification assumptions, sensitivity analysis, evaluation frameworks, monitoring, and post-deployment iteration.
Drive productionization and MLOps practices in collaboration with engineering across distributed data infrastructure.
Track applied research in causal ML, double machine learning, and agentic/LLM systems; translate promising work into production-ready solutions.
Partner with the broader JPMorganChase AI/ML community, model risk, compliance, and peer LOBs to align on standards and amplify firm-wide impact.

Required Qualifications, Capabilities, and Skills

Master's or PhD in Computer Science, Statistics, Economics, Applied Math, Data Science, or a related quantitative field.
3+ years of hands on Machine Learning experience in production environments, with a substantial portion focused on causal inference.
Deep expertise in causal inference methods: potential outcomes framework, propensity score methods, instrumental variables, difference-in-differences, regression discontinuity, synthetic controls, doubly robust and double/debiased ML estimators, and uplift / heterogeneous treatment effect modeling.
Demonstrated experience designing and analyzing experiments (A/B tests, switchback, quasi-experiments) and reasoning carefully from observational data when experimentation is infeasible.
Hands-on experience with LLMs and agentic AI — fine-tuning, RAG pipelines, prompt engineering, and the design and deployment of multi-step / tool-using agents in production.
Strong Python skills; proficiency with causal libraries (DoWhy, EconML, CausalML) alongside PyTorch, scikit-learn, and modern LLM/agent frameworks.
Experience with large-scale data processing: Spark, Hive, SQL.
Proven ability to communicate causal assumptions, limitations, and findings to non-technical stakeholders.

Preferred Qualifications, Capabilities, and Skills

Financial services experience — wealth management, lending, or advisory.
Bayesian and hierarchical modeling; structural causal models; sequential decision-making / contextual bandits.
Experience applying causal reasoning to LLM and agent evaluation — counterfactual eval, off-policy estimation, or treatment-effect framing of agent interventions.