AdamShaar
AI engineer with an M.S. in Computer Science from UIC.
I build production ML systems, and I care most about AI that's correct: I measure models before I trust them, and I build systems that fail safely when they can't verify an answer.
Python · Rust · TypeScript · AWS · RAG · LLM evaluation
// Projects
FisherAI — AI Shelf Card & Excel Generator
Cross-platform Tauri app using YOLO, PaddleOCR, and LayoutLMv3 to turn sale ads into shelf cards automatically, cutting a 2-hour manual task to about 5 minutes.
Tauri (Rust + ORT) · TypeScript · React
Causal Effect of Query Complexity on Product Relevance
End-to-end causal inference pipeline on Amazon’s 793K-row multilingual ESCI dataset, measuring a robust +14.8–16.3% causal lift in exact-match relevance.
Python · EconML · scikit-learn · SBERT
Reliability Beyond Accuracy: Factuality, Uncertainty & Bias in LLMs
Comparative research review of four foundational LLM-reliability papers across factuality, uncertainty calibration, and bias.
Research Paper · University of Illinois Chicago
// Experience
- Junior Solution ArchitectAndrew MorganOct 2026 – Present · Alexandria, Virginia · Remote
- Software DeveloperFisher PrintingJul 2024 – Oct 2026 · Bridgeview, Illinois · Hybrid
// Education
- M.S., Computer ScienceUniversity of Illinois ChicagoAug 2024 – Dec 2025
- B.S., Computer ScienceUniversity of Illinois Chicago2020 – 2024