Applied AI Engineer & Data Scientist
Curiosity led me into AI.
Building became my way of understanding it.
Clayed in Iran, shaped in the U.S.
I recently completed my MS in Data Science at CU Boulder, where my research focused on abductive reasoning and causal inference with language models. My engineering work centers on systems that combine deterministic business logic with LLM reasoning — grounded, auditable, and designed to hold up beyond the demo.
Currently building Lantern Intelligence, an AI financial assistant that grounds every response in real company data, and Lumen, a blackbox LLM evaluation system designed to catch model regressions before they reach users.
Denver, CO
01 —
Selected Projects
01
↗
02
↗
03
↗
04
↗
05
↗
06
↗
Lantern Intelligence
AI accounting assistant — SQL-first architecture, no hallucinated numbers. Runs locally.
Lumen
Blackbox LLM evaluation system — tests any model through inputs and outputs alone.
Abductive Reasoning with LLMs
Why LLMs fail at causal explanation — and how to fix it.
Housing Price Forecasting
Time-series models predicting rent and mortgage costs across Colorado — XGBoost, Elastic Net, 12 models compared.
Multimodal Interview Predictor
Interpretable outcome prediction from speech and text.
Wolfie
Emotion-driven music generation — generative AI applied beyond text.