# Pranav Dulepet

Website: https://pranavdulepet.github.io/
Updated: 2026-06-03

Summary:
Pranav Dulepet works on applied AI at Greenboard in New York City. He recently finished an M.S. in Computer Science at Johns Hopkins CLSP.

Current roles:
- Applied AI team at Greenboard.

Background:
- M.S. in Computer Science from Johns Hopkins University.
- B.S. (Honors) in Computer Science from the University of Maryland.
- Previous work at JHU CLSP, DataTecnica, Apple, Amazon, Fidelity Investments, Capital One, and Evozyne.

Research interests:
- Human-AI collaborative systems
- Large language models
- LLM reliability and evaluation
- NLP and biomedical NLP
- AI agents
- Benchmarking
- AI governance and policy

Primary links:
- Homepage: https://pranavdulepet.github.io/
- Greenboard: https://www.greenboard.com/
- GitHub: https://github.com/pranavdulepet
- LinkedIn: https://www.linkedin.com/in/pranavdulepet
- Twitter/X: https://x.com/pdulepet
- Email: mailto:ps.dulepet@gmail.com

Selected research and writing:
- LLM forecasting for agentic tasks: coming soon
- FAST: Factorizable Attention for Speeding up Transformers: https://arxiv.org/abs/2402.07901
- The Prompt Report: A Systematic Survey of Prompting Techniques: https://arxiv.org/abs/2406.06608/
- Capital One Tech: Similarity Search over Graph Embeddings: https://www.capitalone.com/tech/machine-learning/similarity-search-graph-embeddings/
- Medium: AI safety, alignment, engineering, and policy notes + industry work: https://medium.com/@pdulepet

Selected projects:
- Synopsis AI: https://github.com/pranavdulepet/research-assistant
- agora.: AI meal planning: https://www.agoraai.app/
- LegalAI: legal-document NLP and search: https://github.com/pranavdulepet/legalai
- CollegeRO: https://apps.apple.com/us/app/college-ro/id1577113429

Site notes:
- This site is a simple one-page link index for Pranav's most important references.
- If JavaScript is unavailable, use the metadata, structured data, and this file as the primary machine-readable summary.