Data Domain Architect [multiple Positions Available]

JPMorgan Chase JPMorgan Chase · Banking · Wilmington, DE +1 · Consumer & Community Banking

JPMorgan Chase is seeking a Data Domain Architect to optimize features and AI capabilities for conversational AI products, specifically the Chase Digital Assistant (CDA). The role involves driving NLU model training and optimization, managing intent and entity taxonomy development, and improving training data quality to enhance NLU model performance and intent recognition. The architect will also design analytic frameworks, conduct conversational analysis, and provide linguistic expertise to ML Engineers and Product Managers for new NLP capabilities. This role requires a PhD in Computational Linguistics or related field with 3 years of experience in NLU and AI-powered systems.

What you'd actually do

  1. Optimize features and AI capabilities for the Chase Digital Assistant and other conversational AI products.
  2. Drive NLU model training and optimization for Chase Digital Assistant (CDA), advancing NLU capabilities and conversational AI understanding for improving digital containment within CDA.
  3. Manage intent and entity taxonomy development and align cross-functional teams across Product, Engineering, and Analytics.
  4. Optimize training data sets to improve data quality, NLU model F1 score, and intent recognition rate for CDA.
  5. Partner with Annotation Lead to review and optimize training data and enable meaningful and measurable outcomes.

Skills

Required

  • designing and optimizing Natural Language Understanding (NLU) and AI-powered systems throughout the product lifecycle
  • model training, testing, and evaluation using advanced machine learning and computational linguistics techniques
  • building, testing, and refining language models for AI applications
  • conducting error analysis and quality assurance on datasets to improve model outputs and system reliability
  • creating and maintain taxonomies, and lexical resources to support intent and entity recognition
  • processing and analyzing raw text and datasets to support model development and continuous improvement
  • training, testing, and evaluating model performance using established metrics and methodologies
  • serving as a subject matter expert in linguistics and computational linguistics
  • resolving intent and entity overlaps within taxonomies
  • reviewing and correcting annotation data
  • performing error analysis
  • addressing bugs by collaborating with annotation teams
  • supporting end-to-end AI product development and deployment
  • Notepad++
  • SpaCy
  • Python
  • NLTK
  • Regex
  • Bash
  • Git
  • language modeling frameworks
  • tracking project progress
  • ensuring alignment across cross-functional teams
  • Jira
  • SharePoint

What the JD emphasized

  • PhD in Computational Linguistics, Linguistics, or related field of study plus 3 years of experience
  • two (2) years of experience with the following: designing and optimizing Natural Language Understanding (NLU) and AI-powered systems throughout the product lifecycle, including model training, testing, and evaluation using advanced machine learning and computational linguistics techniques

Other signals

  • NLU model training and optimization
  • Optimize training data sets
  • improve data quality, NLU model F1 score, and intent recognition rate
  • Design extended analytic frameworks and semantic representations to support NLU models
  • linguistic expertise and direction for new NLP capabilities