Shubham Pandey

I’m an AI/ML Engineer with 3+ years of experience building and deploying production-grade Generative AI, NLP, and machine learning solutions. In my current role at Gartner, I build AI-powered research automation systems used by 100+ research analysts, reducing manual analyst effort by approximately 20–30% and shortening research deliverable turnaround by around 30%. I’ve also developed a reusable agentic AI framework that reduced the time required to integrate new AI workflows by approximately 50%, implemented configurable LLM guardrails, and introduced semantic/response caching and prompt optimization that reduced LLM token spend by around 35%. I’ve worked with Python, PyTorch, Hugging Face, LangChain, LangGraph, Azure AI Foundry, Azure OpenAI, AWS, vector databases, RAG, LLM evaluation, and observability.

Previously, I worked on LegalTech solutions involving 5,000+ legal documents, where I built production RAG systems that reduced manual document search time by approximately 50%. I also fine-tuned NLP classification models using PEFT/QLoRA, achieving around 92% F1, and automated six business workflows using LangGraph. Additionally, I contributed to a developer contribution analysis and resume-screening platform that reduced screening time by approximately 60%.

I’m particularly interested in AI Engineer, Generative AI Engineer, ML Engineer, Forward Deployed Engineer, and AI-focused Data Scientist roles. I enjoy taking AI solutions from experimentation and prototyping through evaluation and production deployment, with a strong focus on reliability, scalability, cost optimization, and measurable business impact. I’m especially interested in LLMs, agentic AI, RAG, AI infrastructure, NLP, model evaluation, and building practical AI products in collaboration with product, engineering, and business teams.