Software Engineer III - Ai/ml, Prompt Engineer

JPMorgan Chase JPMorgan Chase · Banking · Mumbai, Maharashtra, India · Commercial & Investment Bank

Software Engineer III focused on Prompt Engineering and Applied ML within a Risk & Controls AI team. The role involves designing and optimizing production-grade prompts, implementing LLM integration patterns (RAG, chain-of-thought), defining evaluation criteria, building agent workflows, and leveraging various databases. Responsibilities include data preparation, feature engineering, monitoring AI/ML workflows, and developing Python scripts. Requires significant experience in ML, applied AI, prompt engineering, and LLM application development.

What you'd actually do

  1. Design production-grade prompts for complex enterprise workflows; test, iterate, and optimize based on outcomes.
  2. Apply familiarity with ML models, including classification, NLP, and transformer-based architectures. Implement LLM integration patterns such as retrieval-augmented generation (RAG), chain-of-thought prompting, and response validation.
  3. Define and execute prompt/model evaluation criteria (accuracy, consistency, hallucination rate, policy adherence).
  4. Design and run offline and online experiments to improve prompts and tune model performance.
  5. Implement guardrails, safety filters, and fallback strategies in production AI/ML workflows. Use ML frameworks such as PyTorch, TensorFlow, scikit-learn, and/or Hugging Face Transformers.

Skills

Required

  • BS/BA degree in Computer Science, Engineering, Machine Learning, Data Science, Statistics, or equivalent experience
  • 5+ years of hands-on experience in machine learning, applied AI, or prompt engineering in production environments
  • 5+ years of experience with Python and ML/AI tooling (model development, evaluation, and deployment)
  • 3+ years of experience with LLM application development, prompt engineering, or NLP systems
  • 1+ years of experience with experimentation frameworks, A/B testing, and production monitoring of AI features

Nice to have

  • Experience with responsible AI practices and model governance
  • Experience in financial services, technology risk, or controls-oriented environments
  • Relevant ML/AI certifications (e.g., AWS ML Specialty, Google Professional ML Engineer)
  • Understanding of enterprise risk, controls, and auditability expectations

What the JD emphasized

  • production-grade
  • production environments
  • production monitoring
  • enterprise workflows
  • controls outcomes
  • risk and control outcomes
  • controls stakeholders
  • enterprise risk

Other signals

  • production-grade AI
  • prompt engineering
  • applied machine learning
  • LLM integration patterns
  • agent workflows