New York, New York, United States
Developed a redesign of a network loss tracking client, adjusting the data ingestion and evaluation processes for more efficient faulty device detection leading to reduced ticket creation times by, on average, 15 minutes.
Integrated various frontend components including dashboarding and sharing features in internal tooling to allow greater efficiency of on-call knowledge sharing and reducing widespread duplication of knowledge.
Integrated backend storage services in Rust for a feature flagging service to be deployed and improve internal version changes in over 3 supporting services.
Designed and implemented a new signal processing algorithm for light levels from device fibers, identifying potential issue locations in WAN network device connections with accuracy to 90% of all runs.
Developed LLM based workflows to remediate live network faults with LLMs driving responses and pushing forth activities to be conducted through email, teams messages, and custom activities, improving remediation from 24 hours to 4 hours.
Designed and deployed an LLM evaluation pipeline running automatically on prompt changes to in-production AI agents mitigating network operations issues, ensuring >90% compliance with expected and actual response outputs.
