Neural Network Architect

About the Role

We are building a new AI function to modernize how our metals operations generate, connect, and act on data. The **Neural Network Architect** will lead the design and implementation of a **unified data and AI infrastructure** that links our Enterprise Resource Planning (ERP), shop floor process/control systems (SCADA, PLC, MES/historian data), Laboratory Information Management Systems (LIMS), cost accounting systems, and Customer Relationship Management (CRM) platforms into a single, coherent architecture.

This is a foundational, cross-functional role. The person in this seat will not just design neural network models, but they will architect the data backbone that makes reliable, production-grade AI possible across the enterprise: from predictive quality control and yield optimization on the shop floor, to demand forecasting that links CRM and production data, to cost-to-serve models that tie financials to real-time operations.

SCOPE:

While "Neural Network Architect" is the title, this role is as much about systems integration, data engineering, and information architecture** as it is about neural network design. Candidates should be equally comfortable designing a data pipeline between a PLC historian and an ERP as they are designing or evaluating a deep learning model. We are not looking for a pure research scientist, rather we are looking for someone who can turn fragmented industrial data into an infrastructure that AI systems (including neural networks) can reliably run on.

Key Responsibilities

Unified Infrastructure Design

• Architect a scalable, secure data platform that integrates ERP, shop floor/process control systems, LIMS, cost/financial systems, and CRM into a shared data layer.
• Define data models & schemes to link above data with data security and pathways for statistical models to be integrated
• Establish real-time and batch data pipelines connecting operational technology (OT) systems on the shop floor with information technology (IT) systems (ERP, CRM, finance).

Neural Network & AI Model Development

• Design, train, and deploy neural network and other ML models on top of the unified data infrastructure for use cases such as predictive maintenance, quality/defect prediction, yield optimization, demand forecasting, and cost anomaly detection.
• Evaluate where deep learning is the right tool versus simpler statistical or rules-based approaches, given industrial data constraints (sparsity, sensor noise, labeling challenges).
• Build MLOps practices for model versioning, monitoring, and retraining as plant and market conditions evolve.

Systems Integration & Governance

• Partner with IT, OT/automation engineers, lab managers, finance, and sales operations to map data sources, ownership, and access rules across departments.
• Establish data governance standards: data quality, lineage, security, and access control across sensitive systems (cost data, customer data, process parameters).
• Select and implement integration tooling (APIs, middleware, historians, data buses) appropriate for connecting legacy industrial systems with modern cloud/on-prem AI infrastructure.

Leadership & Strategy

• Serve as the technical authority on how AI infrastructure should be built as the company's AI department scales.
• Advise leadership on architecture trade-offs, build-vs-buy decisions, and sequencing of integration work (e.g., which systems to connect first for fastest ROI).
• Document architecture decisions and mentor future data engineers/ML engineers hired into the department.

## Required Qualifications

• Bachelor's or Master's degree in Computer Science, Data Engineering, Electrical Engineering, Industrial Engineering, or related field (PhD a plus but not required).
• 7+ years of experience in data architecture, data engineering, or ML/AI engineering, including at least 3 years working with industrial or manufacturing data (OT/IT integration experience strongly preferred).
• Demonstrated experience designing and deploying neural networks or other ML models in production (not just research/prototype settings).
• Hands-on experience integrating disparate enterprise systems a" direct experience with ERP platforms (SAP, Oracle, Infor, etc.), MES/historian systems (e.g., OSIsoft PI, Ignition, Wonderware), and CRM platforms (Salesforce, Dynamics, etc.) is highly valued.
• Strong data engineering skills: SQL, Python, data pipeline tools (Airflow, dbt, Kafka, or similar), and cloud or on-prem data platforms (Azure, AWS, GCP, Snowflake, Databricks).
• Working knowledge of ML/DL frameworks (PyTorch, TensorFlow) and MLOps tooling.
• Understanding of industrial control systems architecture (SCADA, PLC, Purdue Model / ISA-95 levels) and the security considerations of bridging OT and IT networks.

## Preferred Qualifications

• Prior experience in metals, mining, or heavy manufacturing industries.
• Familiarity with LIMS platforms and quality/metallurgical data structures (chemistry, mechanical properties, heat treatment data).
• Experience with cost accounting or ERP-based financial data models (standard costing, activity-based costing).
• Experience building a data/AI function from the ground up in a traditionally non-digital organization.
• Knowledge of relevant industrial data standards (ISA-95, MTConnect, OPC UA).

Success Metrics for This Role (First 18 Months)

• A functioning, documented unified data architecture connecting at least ERP + shop floor + LIMS as a first milestone, with CRM and cost systems integrated in subsequent phases.
• At least one production AI/ML use case (e.g., a quality prediction or yield model) running on top of the new infrastructure.
• Clear data governance and access framework adopted across departments.
• A scalable foundation that allows the AI department to add additional models and use cases without re-architecting from scratch.

All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, or status as a protected veteran.

With more than 120 operations and approximately 20,000 employees worldwide, Precision Castparts Corp. is the market leader in manufacturing large, complex structural investment castings, airfoil castings, forged components, aerostructures and highly engineered, critical fasteners for aerospace applications. In addition, we are the leading producer of airfoil castings for the industrial gas turbine market. We also manufacture extruded seamless pipe, fittings, and forgings for power generation and oil & gas applications; commercial and military airframe aerostructures; and metal alloys and other materials for the casting and forging industries. With such critical applications, we insist on quality and dependability - not just in the materials and products we make, but in the people we recruit.

PCC is relentless in its dedication to being a high-quality, low-cost and on-time producer; delivering the highest value to its customers while continually pursuing strategic, profitable growth.

In 2016, Berkshire Hathaway, led by Chairman and CEO Warren E. Buffett, acquired Precision Castparts Corp.