Principal ML Engineer

Wells Fargo Bank•Irving, TX

About The Position

COO Technology operates at the center of the enterprise, enabling the Chief Operating Office to execute with precision, resilience, and confidence. Within it, the Cognitive AI Solutions Team is building the platform that powers the next generation of agentic AI and process intelligence for the firm: a system that automatically discovers how work actually gets done; from multimodal signals including video, documents, clickstream data, and system logs and turns that understanding into AI agents that improve it. This is not a lab. The platform serves multiple lines of business across one of the largest, most regulated operational environments in the world, and every capability we build is designed to be reused across many use cases. We are seeking a Principal ML Engineer to be the technical anchor of the platform itself; a hands-on architect who has built a product or platform that serves many use cases, not a single model or a proof of concept. You will own the hardest technical problems in the system: automating process discovery at enterprise scale, extracting structured insight from multimodal data, and designing the reusable components; agent harnesses, knowledge graphs, evaluation frameworks; that let delivery teams build agentic solutions in weeks instead of quarters. You will set technical direction and then build it yourself. The RIGHT candidate has shipped platforms that other engineers build on, has built and deployed AI agents in production, and can move fluidly between systems architecture, ML modeling, and code.

Requirements

7+ years of Engineering experience, or equivalent demonstrated through one or a combination of the following: work experience, training, military experience, education.
3+ years of experience with clustering algorithms and unsupervised learning techniques for pattern discovery.
2+ years of experience working with knowledge graphs or graph-based ML techniques.
5+ years of experience with Python with experience in ML frameworks.

Nice To Haves

Experience with agentic data retrieval and analysis at the Enterprise Level.
3+ years of experience building entity resolution, deduplication, or record linkage systems at scale.
Expertise in NLP techniques for semantic similarity, text clustering, and information extraction.
Experience with graph databases.
Background in process mining, conformance checking, or business process analysis.
Experience with LLMs and embedding models for semantic similarity.
Experience building ML pipelines using MLOps best practices.
Experience with cloud computing platforms.
Experience with distributed computing frameworks.
Knowledge of containerization and orchestration technologies.
Experience in financial services or operations domains.
Excellent communication skills across technical and non-technical audiences.
Advanced degree (M.S. or Ph.D.) in Computer Science, Machine Learning, or related field.

Responsibilities

Architect the platform. Own the end-to-end architecture of a process-intelligence platform designed to serve multiple business domains — defining the abstractions, APIs, and reusable components that make agentic solution-building repeatable and scalable.
Automate process discovery. Design and build ML systems that reconstruct how enterprise processes actually execute — mining patterns from clickstream data, documents, videos, and logs; reconciling fragmented entities; and extracting canonical execution paths at scale.
Build multimodal understanding pipelines. Apply foundation models and state-of-the-art techniques to extract structured insight from unstructured, multimodal sources — video recordings of work, procedure documents, transcripts, and transactional logs.
Engineer agentic systems. Design agent architectures, orchestration frameworks, tool interfaces, and memory/knowledge-graph layers that turn discovered process intelligence into reliable, autonomous AI agents.
Enrich enterprise knowledge graphs. Build entity resolution, intelligent clustering, and conformity-analysis capabilities that measure real-world execution against documented procedures and continuously learn from new data.
Set the engineering bar. Establish standards for evaluation, observability, guardrails, and production reliability across the platform; mentor senior engineers and raise the technical ceiling of the team.
Own delivery of the hardest problems. Personally prototype, productionize, and ship the most complex components — and be accountable for their performance in a regulated, enterprise-scale environment.