🎓 MS CS (AI) @ USC (May 2024 to May 2026) , BS Softwware Engineering @ UArizona (August 2019 to December 2023)| 💻 Software Engineer building ML systems, data pipelines, and agentic LLM workflows | 🛡️ Interested in LLM and agentic AI for defense and financial applications
📍 Los Angeles 🌴 | 🏠 Seattle 🌧️

• 🤖 Agentic AI systems that can plan, reason, and execute reliably in real environments
• 🌌 LLM + knowledge graph workflows for better retrieval, grounding, and decision support
• 🧱 Data pipelines and analytics tooling that make metrics trustworthy and actionable
• 🧪 Productionization: containers, reproducibility, observability, and clean interfaces

• 🧠 Built an ML tool at USC ISI integrating knowledge graphs and LLMs to deliver AI-driven insights for government-sponsored research, including agentic workflows and Dockerized services using Ollama for inference.
• 📈 Pinata Cloud: analyzed 3+ years of sales and product usage data (Snowflake, MongoDB) and delivered executive insights that supported a 10% pricing increase. Won a hackathon by building an AI content classification feature (Python, Flask, Google Cloud Vision) that shipped to production. Built utilization dashboards in Mode and Hex that cut leadership analysis time by ~20%.
• 🛠️ Raytheon Missiles & Defense: improved telemetry transformations by integrating Python preprocessing into Azure DevOps pipelines, and automated extraction workflows, contributing to large annual savings. Built real-time telemetry visualization using Grafana, InfluxDB, and JavaScript to speed debugging and improve visibility.
• 🧾 Costco: built real-time Power BI dashboards, optimized large SQL datasets, and integrated Azure Databricks with Python and NumPy into ETL workflows to improve forecasting and inventory operations.

Languages

Data + ML

Platforms + Tools

Visualization

LLM stack

Exploring how to build agentic systems for complex settings, with emphasis on:

• ✅ data quality and governance
• 📏 evaluation and reliability
• 🔁 scalable pipelines for analytics and model feedback loops

• LinkedIn: https://linkedin.com/in/harrison-pearl-ai
• Email: hpearl@usc.edu