Kermit Troy Berry

Hybrid spearheaded Splunk Observability integration for McLaren Racing: designed 100kHz Open Telemetry ingestion pipelines and custom ITSI dashboards. Developed advanced SPL correlation searches to detect anomalies, reducing mean time to resolution by 25%. Extended Splunk Phantom playbooks with PyTorch NLP for structured log triage, boosting SOC throughput by 30%. Automated dashboard provisioning via Splunk RESTful API and Terraform ensuring staging/production consistency. Trained cross-functional teams on ML Toolkit and Observability best practices to democratize insights.

Led phase 2 rollout of C3.ai Energy Management across 600 analytics modules. Developed and deployed closed loop HVAC control models using XGBoost, MLflow, and Python on Linux servers, delivering a 15% incremental energy savings uplift. Orchestrated high throughput Apache Kafka and MQTT pipelines ingesting 1M events. Built CI/CD workflows with Jenkins and C3 CLI automating versioning and validation, reducing deployment time by 60%. Monitored model drift and system health via Prometheus and Grafana, establishing proactive rollback procedures to maintain 99.9% uptime.

Partnered with IGS Energy to deploy Snowflake Cortex ML built MLOps pipelines and SQL-based feature engineering boosting demand forecast accuracy by 22%. Automated ETL/feature workflows with Streams & Tasks, halving model retraining latency. Tuned clustering keys and materialized views to accelerate queries by 40% without extra compute. Enforced production model deployments as RESTful APIs using Python and MLflow. Developed custom JavaScript and Python UDFs for anomaly detection and bespoke transformations.