Dave Hemler

Senior Director

Industry Labs

Machine Learning, Workflow Automation, Data Cleaning & Preparation, Exploratory Analysis, Forecasting & Optimization, Dashboards

Full Overview

We build relationships with local companies and support programming that aligns campus resources and industry expertise to unlock value and imagine what is possible.

We engage faculty and students with companies located around the South Bend region through our AI Community of Practice and student-delivered projects.

Example Projects

HR & Workforce

Multilingual HR Assistant: A workforce with a significant number of non-English-speaking employees had difficulty accessing and interpreting HR handbooks and policy documents. The project built a chatbot that lets employees query HR content in 4 different languages, reducing the HR team's workload.

Automated Apprentice Progress Tracking: Performance, training, and retention data for a structured apprentice program were scattered across multiple systems, forcing supervisors to manually compile monthly updates. LLMs were used to architect data-consolidation strategies and to generate a custom dashboard with supervisor-ready progress metrics and summaries.

Job Code Rationalization: A large organization managing hundreds of job descriptions in its HR system faced challenges with hierarchical role management and skills/certification tracking. The project uses AI to align roles with industry standards and builds an agent that helps HR directors add new job descriptions to the rationalized library sustainably.

Finance & Accounting

Month-End Financial Statement Analysis: An accounting team with limited capacity was unable to conduct thorough month-to-month financial analysis and variance commentary. The project developed an AI-powered process that aggregates general ledger transactions into basic financial documents and produces a structured summary for leadership. The project also explored an LLM chatbot that allows the C-suite to further query financial statement trends and insights over time.

A/P Document Cross-Reference Automation: Discrepancies between invoices, purchase orders, and receiving documents were creating financial accuracy issues and requiring manual review. An agent will automatically flag mismatches and escalate exceptions for human intervention.

Operations & Supply Chain

Accessories & Parts Demand Forecasting: An organization with several years of sales history was struggling to achieve acceptable forecast accuracy across a diverse parts portfolio, with ML models in Azure performing well for some categories but falling short on others. The project provided hands-on model development with both traditional data science, Python-driven analysis, and a Claude-developed forecasting skill that reduced forecasting time from weeks to hours and provided inventory policy recommendations.
Network Optimization: A large organization operating national production and distribution facilities had accumulated significant inefficiencies in how work was allocated and distributed across sites, with recent investments in certain locations not being fully utilized. Notre Dame data scientists built a cost-to-serve optimization model that analyzed production, distribution, and site capability data to recommend the most efficient network configuration. The project resulted in a reduction of nodes from 22 to 13 and multi-million dollar cost savings.
Automated Supplier Scorecard: Procurement teams manually pulled and analyzed data to generate supplier performance metrics across lead time, quality, and other dimensions. The project will automate the scorecard generation process, freeing buyers to focus on higher-value, strategic work.
High-Value Procurement Agent: Back-and-forth communication between customers and suppliers on large contracts was causing quote expirations and pricing inaccuracies due to commodity price volatility. An AI agent will be designed to accelerate the quoting cycle and audit quoted vs. billed amounts in near real time.

Sales & Business Development

Inbound Lead Triage Agent: Sales leads submitted through a web form required manual review, CRM entry, and legitimacy screening before advancing. The project built an agent to analyze submissions, conduct brief background research, and prioritize leads. Furthermore, another agent was built to draft initial response emails and put them into a signed salesperson's Outlook draft.

Market Research & Outreach Agent: A business development team manually searched for target companies and decision-makers when entering new markets, a slow, inconsistent process. An AI agent will be built to ingest internal product and application data, identify prospect companies, surface relevant contacts, and evolve toward automated personalized outreach.

IT / Data Access

Natural Language Database Query: Non-technical staff were unable to efficiently query operational data in response to customer inquiries, requiring complex manual queries or analyst involvement. The project built a chatbot interface on top of a data warehouse, allowing
plain-language questions to return structured query results. Aside from using MCP, this project also focused on heavy contextualization of the data warehouse and its ontology.

Knowledge Management

Institutional Knowledge Capture Agent: Critical operational knowledge was concentrated in a small number of experienced staff, creating organizational risk. The project leveraged Murray Mentor, a third-party tool for tacit knowledge capture and management, to capture experienced staff knowledge and distribute it via an on-demand chatbot to less tenured employees.

Project & Program Management

AI-Assisted Task Prioritization: Engineers consistently struggled to prioritize non-critical tasks after their primary workload was complete, sometimes blocking critical project paths. The project explored how AI could surface low-visibility tasks that carry hidden organizational importance within Microsoft Planner.

IP Related Considerations

Our approach to intellectual property (IP) in collaborations is flexible and aligned with university policies and sponsor agreements. Industry partners typically retain ownership of their background IP. For project-generated IP, ownership depends on factors such as funding source, contractual agreements, and contribution to invention (authorship). In many applied student-led projects, deliverables are structured to ensure that partners can directly use outcomes (e.g., reports, models, dashboards), while the university may retain rights for educational and research purposes. When required, formal agreements are used to clearly define IP ownership and usage rights.

Student Level

Undergraduates, Masters, Mix

Budget

Fall, Spring, Summer

Typical Team Size

1-2, 3-5

Terms Available

Fall, Spring, Summer

Delivery Model

Hybrid (Faculty + Students)

Interested in engaging in a project?