~70% less manual work preserving engineering knowledge from a retiring system

Years of machine-configuration knowledge lived in XML documents inside a system scheduled for retirement. Re-creating it by hand as decision logic and configuration rules in the replacement platform meant months of expert time.

• Co-built, with the client team, an LLM/NLP pipeline that reads the legacy XML and generates the decision logic and configuration rules the new platform needs.
• Built the migration interface on the target platform and deployed it with Terraform on AKS.
• In parallel, built and demonstrated a Databricks-centric MLOps framework with example workflows, and a supplier-risk assessment agent for the procurement function.

• 80–90% pipeline accuracy with ~70% less manual workload (joint client–UTIS team result).
• Migration interface deployed and demonstrated; production rollout in progress at time of writing.
• The risk agent is projected to cut manual review time by ~35% once adopted.