Abhigya Verma

About Abhigya Verma

I am a Machine Learning Engineer at ServiceNow, specializing in building and scaling large language models (LLMs) for enterprise applications. My work spans synthetic data generation, multimodal reasoning, and safety alignment, where I have designed and deployed pipelines that power real-world products.At ServiceNow, I contributed to the NowLLM initiative by generating high-quality multimodal datasets (image-to-text, reasoning, Q&A) and enhancing safety alignment for content moderation and red-teaming use cases. Previously, I conducted research at NAIST Japan, developing a real-time ASR system for Hindi, and interned at Microsoft, where I enhanced the Outlook Search Assistant with intelligent query suggestions.My academic and research contributions include peer-reviewed publications on synthetic data generation and multimodal benchmarks, such as GRAFT, which evaluates instruction-following and visual-textual alignment in LLMs. Recognized with awards like the UTG Llama Award (ServiceNow) and a GHCI Student Scholar, I bring both technical depth and community leadership experience.I am passionate about advancing trustworthy, efficient, and scalable GenAI systems—bridging cutting-edge research with enterprise-scale engineering to make AI safer, more robust, and impactful.

Experience

1. Machine Learning Engineer

   ServiceNow

   Feb 2025 — Present · Hyderabad, IN

   Generated image-to-text and reasoning text data for NowLLM multimodel scope.• Created evaluation datasets for model safety and security accuracy.• Built end-to-end pipelines to evaluate model metrics for LLMs.

Education

• INDIRA GANDHI DELHI TECHNICAL UNIVERSITY FOR WOMEN

  B.Tech, IT

  2019 — 2023

• Oxford Senior Secondary School - India

  High School

  2007 — 2017

• Oxford Senior Secondary School - India

  High School, Medical Science