Piyush Ghai
Staff Applied Scientist, Relativity | Ex-AWS AI | Stanford GSB Lead
- Role
- Staff Applied Scientist at Relativity
- Location
- San Francisco, CA, US
- LinkedIn followers
- 500 followers
About Piyush Ghai
I am a Staff Applied Scientist at Relativity, where I lead applied research teams in building AI systems that tackle high-stakes problems in privacy, compliance, and legal domains. My work spans NLP, deep learning, and generative AI (LLMs), with a focus on bridging research and production to make large-scale models reliable, explainable, and impactful. Previously, I contributed to distributed training at AWS SageMaker, and I’m passionate about advancing AI systems that are both technically rigorous and enterprise-ready.
Experience
Staff Applied Scientist
Aug 2021 — Present · San Francisco, CA, US
I am a Staff Applied Scientist at Relativity, where I lead applied research teams at the intersection of machine learning, natural language processing, and generative AI. I joined through the acquisition of Text IQ, and since then I’ve been driving innovations that help enterprises navigate privacy, compliance, and legal challenges.Key initiatives I’ve led include:Precision Enhancer – Built graph + text hybrid models to detect communication anomalies with high precision/recall, improving noise resilience in enterprise-scale data.PII Detection – Designed BERT-based representation learning with SimCSE pretraining and fine-tuning for sensitive data classification.aiR for Privilege (PrivIQ) – Designed and delivered an end-to-end system for detecting privileged documents and classifying privilege types, with built-in guardrails for reasoning, citations, and hallucination mitigation.RAG Search – Developed POCs for retrieval-augmented generation pipelines combining hybrid search, re-ranking, and LLMs to power trustworthy document insights.LLM Evaluations – Leading efforts to define reliable evaluation methods for LLMs, combining rubric-based judging, alignment across evaluators, and synthetic dataset generation.My work is grounded in tackling fundamental challenges with LLMs—hallucinations, reproducibility, long-context reasoning—and building systems that make these models reliable for enterprise use.
Education
The Ohio State University
Master’s Degree, Computer Science
2016 — 2018
Stanford University Graduate School of Business
Stanford Lead
Delhi Public School - R. K. Puram
High School
2008 — 2010
New Era Public School - India
High School
1996 — 2008
Netaji Subhas Institute of Technology
Bachelor's Degree, B.E ( Information Technology )
2010 — 2014
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.