Abdul Ahad

About Abdul Ahad

I build AI systems the way an engineer–scientist should: by questioning assumptions, running experiments, and shipping what actually works.With a foundation in Physics and advanced training in Artificial Intelligence & Machine Learning, I operate at the intersection of machine learning research and production-grade engineering. I don’t treat models as abstract benchmarks—I treat them as systems that need to be understood, stress-tested, and made reliable in the real world.My hands-on work centers on deep learning with PyTorch, where I design controlled experiments to study training dynamics, optimization behavior, activation choices, and failure modes. I focus on why models behave the way they do, because that insight leads to systems that are easier to debug, more stable to scale, and safer to deploy.Beyond model internals, I build applied NLP and Retrieval-Augmented Generation (RAG) pipelines, paying close attention to data structure, retrieval quality, latency, and system trade-offs. I enjoy creating lean, extensible prototypes—research-driven enough to measure and iterate, yet practical enough to evolve into real products.What I bring to a team is clarity: the ability to connect theory to implementation, research to engineering, and experimentation to shipped outcomes.I’m seeking AI/ML Engineer or AI Systems Engineer roles in fast-moving environments where ownership matters, learning is continuous, and deep technical thinking is expected—not optional.

Experience

1. Associate Prompt Engineer

   NVIDIA

   Apr 2026 — Present · Pune City, IN

   Deputed at NVIDIA as part of the LLM Data & AI team.Contributing to NVIDIA\'s generative AI initiatives by creating and evaluating high-quality training data for large-scale language and multimodal AI models.Developing conversational, generative, and descriptive prompt-response pairs for AI model trainingGenerating image captions, video descriptions, and video event captions for multimodal AI systemsSummarizing video content and writing detailed descriptions with and without audio contextEvaluating and scoring human and AI-generated prompt-response pairs across multiple quality parametersComparing, rating, and refining multiple AI-generated responses to improve output quality and consistencyCollaborating with cross-functional teams to maintain high standards for AI training data

Education

• Jamia Millia Islamia

  MSc, Artificial Intelligence & Machine Learning

• Jamia Millia Islamia

  Bachelor of Science - BS, Physics

• Wilsonia College

  High School Certificate (10th), PCM + Computer Science

• Wilsonia College

  11th-12th, PCM+Computer Application