Anuj Mittal

Ai Trainer Mathematics @Pareto AI

IN

Over the past two years at Pareto.AI, I’ve spearheaded a variety of AI alignment and training initiatives: in Project Barium, I crafted and evaluated original math puzzles to sharpen model reasoning; Project Hafnium saw me annotating and curating rich problem sets for improved model comprehension; Project Mercury had me transcribing complex textbook questions into machine‑readable form; Project Lawrencium involved developing nuanced political prompts for fair and accurate discourse; with Multi‑Modal Rubrics and Project Beryl, I designed and rigorously tested visual‑question‑answer tasks to advance a model’s multi‑modal understanding; and in Project Richterite, I created and compared creative‑writing prompts to help models generate more engaging, context‑appropriate text. Together, these projects highlight my strengths in prompt engineering, content development, and rigorous evaluation.

IIT Roorkee

Master’s Degree