About
I'm a Senior Software Engineer at AT&T, where I work on the Business AI Foundry. I architect and develop scalable Agentic AI-powered workflows across our Business Internet services. My work sits at the intersection of backend systems, agentic orchestration, and applied ML. I lead technical delivery for a six-engineer team, and I'm the organization's go-to person for AI evaluation and observability, supporting 15+ teams.
I'm currently pursuing an M.S. in Computer Science at Georgia Tech, expected 2027.
I graduated from Rice University with a BA in Computer Science and a BS in Mathematics. While there, I served as President of Hanszen College, one of Rice's eleven residential colleges, leading our elected student government and advocating for students across campus.
I formerly ran my own tutoring business, working with college, high school, and middle school students in Mathematics, Computer Science, Physics, Chemistry, and Latin.
Outside of work, I play soccer, water polo, ultimate frisbee, and table tennis. I play violin and piano, produce music, and do transcription. I also speedcube and juggle — here's a video of me solving a Rubik's cube blindfolded.
In July 2026, I won the game show The 1% Club (Season 3 Episode 9) "I'm Just Dumb."
Experience
- Led architecture and development of a production AI-powered customer support platform serving 2.5M business customers and processing 300K+ incidents and requests monthly; projected to save ~$40M annually.
- Architected agentic workflows in Python and LangGraph that analyze network diagnostics, customer feedback, and operational data to identify root causes, resolve issues, and escalate exceptional cases.
- Engineered fault-tolerant workflows using LangGraph checkpointing, persistent state, and retry and recovery logic for reliable execution of multi-step enterprise processes.
- Led an architectural rewrite of a platform's core business logic using agentic workflow patterns, improving test-plan accuracy from ~80% to 98% — contributing to ~$100M in projected business value.
- Designed LLM evaluation, regression testing, and observability frameworks with LangSmith; primary AI evaluation SME supporting 15+ teams and 100+ technical stakeholders, delivering training, consulting, and documentation on AI engineering best practices.
- Built integrations with ServiceNow, enterprise ticketing systems, and Microsoft Graph for agent-driven workflow execution, plus full-stack features for an AI workflow platform serving 11,000+ internal users.
- Led technical delivery across a six-engineer team — architecture, planning, code review, and releases; mentored 10+ engineers, presented AI solutions to CTO-level leadership, and helped scale the org from 7 to 100+ people.
- Built an internal Territory Visualization Tool (React/TS, PHP/Mezzio, MySQL) that eliminated sensitive data exports to vendors and improved sales territory precision 600x.
- Created analytics and reporting capabilities to help balance sales representatives' workloads.
- Automated user management, approval workflows, security policy enforcement, and reporting for internal financial systems using UiPath RPA.
- Produced network architecture diagrams, video demos, and software documentation.
Skills
- Languages
- Python, Java, JavaScript / TypeScript, SQL
- Backend
- REST / SOAP APIs, distributed systems, concurrent systems, OOP design patterns
- Frameworks
- FastAPI, Angular, React, LangGraph, LangSmith, LangChain, LaunchDarkly
- Databases
- PostgreSQL, MongoDB, MySQL
- AI / ML
- RAG, vector embeddings, agentic orchestration, LLM and agent evaluation, prompt engineering, observability
- DevOps
- Docker, Kubernetes, Azure DevOps, Git
Competitions
- Developed machine learning models to predict social network connections using TensorFlow and Python.