Jordan Randleman

Developed full-stack LLM assistant experiences for a benefits platform serving insurance brokers, HR admins, and 3.7M+ employees. Key Contributions: ● Hardened production benefits-chatbot behavior across company-specific benefits contexts using MLflow-backed prompt versioning, tool-call routing, dynamic link whitelisting, and internal tool/source-flow guardrails ● Designed Broker AI backend agentic architecture using skills-based retrieval, dynamic tool loading, and progressive skill summarization to preserve prompt-cache efficiency ● Built Figma-aligned Broker AI frontend across sidebar, full-page, and dashboard workflows with conversation history, client mentions, upload affordances, message actions, rich editable inputs, and independent draft state

Developed AI tutoring infrastructure powering a platform with 5,000+ monthly active users for a pre-seed startup. Key Contributions: ● Built a multi-model agentic LLM tutoring evaluation framework scoring performance across 27 pedagogical dimensions, improving tutor quality up to 65% over baseline models and achieving 94.37% math accuracy, inspired by Google’s LearnLM research ● Built and deployed 50+ backend TypeScript microservices spanning conversation management, media uploads, problem sets, error reporting, usage metrics, and onboarding flows ● Led migration of core backend to Next.js Route Handlers + AWS Lambda, improving deployment speed, reliability, and codebase maintainability ● Improved the Python code execution's Docker Lambda sandbox for math problems, increasing execution reliability and cutting parsing-related failures by 61% ● Optimized TypeScript Manim video rendering and HLS streaming pipeline with parallelization and caching, reducing animation processing times by 70% by improving the open-source Manim library ● Implemented comprehensive monitoring and analytics dashboards to track 80+ CloudWatch model errors, latency events, and production health: reducing debugging friction and improving reliability

Taught CSEN 171: Programming Languages, an upper-division course exploring the syntax, semantics, and fundamental concepts of programming languages across four paradigms: ● Imperative Programming (C, Pascal) – Procedural abstraction and direct memory manipulation ● Object-Oriented Programming (C++, Java, Smalltalk) – Encapsulation, inheritance, and polymorphism in an object-centric approach ● Functional Programming (ML, Scheme, Scala) – Pure functions, recursion, and higher-order functions for declarative programming ● Logic Programming (Prolog) – Computation through deduction via facts and inference rules Through hands-on coding practicums and theoretical analysis, students gain a deep understanding of how different programming paradigms shape software development and computational thinking. Professor Rating: 4.7/5.0

Created an appointment management system for hospitals and patients in Senegal, in association with Sayap Africa. I helped architect and develop the hospital's appointment management system using Flutter and Firebase, as well as design our database table schemas.

● Fine-tuned Meta’s Llama 2 LLM on internal tooling data to enhance Firstup’s developer experience ● Built an LLM Dispatch Service: Flask web server wraps OpenAI APIs with a price prediction algorithm for automated model routing, JSON generation improvements for enhanced hashmap synthesis, and SQL logging

● Built an internal web portal using C# and ASP.NET to survey Azure network workload simulations ● Visualized network simulation data using Azure Storage Table and Azure Blob Storage UIs ● Monitored simulation system health with inter-process communication logs sent over Azure Storage Queues Received a full-time return offer, but declined to go to grad school and keep learning :)

● Overhauled Data Analytics Suite: updated 77 concurrent Node.js scripts (25 new, 52 updated) for mobile app ● Built an internal web portal to monitor system health: monitors the activity of 40,000+ weather stations ● Integrated IoT APIs: EC2 Server controls 5 Honeywell smart thermostat settings for Wynd’s air purifier

GPA: 3.91

GPA: 3.95 Minor in Mathematics