Has Arnav Raj submitted work to NeurIPS?

Yes, co-author on a NeurIPS 2025 Datasets & Benchmarks submission (KG-QAGen) for structured multi-hop question generation and long-context evaluation.

How to find Arnav Raj on Google Scholar?

Search 'Arnav Raj Google Scholar' or follow the Scholar link on the site; profile lists preprints and submissions.

About

Q: What research areas is Arnav Raj focusing on?

LLM evaluation (reasoning reliability, retrieval grounding), retrieval augmented generation, reinforcement learning exploration, and systems for ML (performance & observability).

Q: What is Arnav Raj's academic background at IIT Delhi?

Dual Degree (B.Tech + M.Tech) Computer Science & Engineering focusing on computer architecture, operating systems, AI systems reliability.

Q: What competitive achievements does he have?

JEE Advanced AIR 1158, KVPY SX Fellow, olympiad stage performances, Codeforces Expert rating.

Arnav is a fourth-year dual degree student at IIT Delhi pursuing B.Tech and M.Tech in Computer Science & Engineering with a focus on AI safety and LLM reliability. Currently working as an AI Training Data Research Intern at Abundant AI, he designs challenging tasks that expose failure modes in state-of-the-art language models, contributing to datasets used by leading AI labs and Fortune 500 companies. Beyond academics, Arnav co-founded the AI Safety Club at IIT Delhi, completed structured alignment training through BlueDot Impact, and enjoys exploring new places and culinary experiences.

Research-wise, Arnav recently submitted solo work to ICLR 2025 on detecting hallucinations via hyperbolic geometry (achieving 87.5% accuracy, 0.937 AUROC). His research internships at Harvard University (Edge Computing Lab) and Georgia Tech (FSI Lab) focused on LLM benchmarking for RTL code generation and long-context evaluation respectively. He co-authored KG-MuLQA, submitted to ACL 2025, which provides 20,139+ multi-hop QA pairs for structured reasoning evaluation. Ranked All India 1158 in JEE Advanced (top 0.1%) and a KVPY SX Fellow, Arnav maintains an active research profile exploring the intersection of AI safety, interpretability, and reliable evaluation systems.

If you searched for terms like “Arnav Raj IIT Delhi”, “Arnav Raj AI Safety”, “Arnav Raj ICLR”, or “Arnav Raj hallucination detection”—this page consolidates those identities: researcher focused on making LLMs more reliable, interpretable, and safe.

Download Full Resume (PDF)

Education

Indian Institute of Technology Delhi
B.Tech and M.Tech in Computer Science & Engineering, 2022-2027
Mess Secretary, Zanskar Hostel
Leadership & Operations, Jun 2024 – May 2025
Senior Editor, Tech Ambit (Pan-IIT Magazine)
Editorial & Tech Strategy, 2023 – 2025

FAQ

What research areas is Arnav Raj focusing on?

AI safety, hallucination detection in LLMs, RLHF/RL training data design, long-context evaluation, model interpretability, and retrieval-augmented generation. He's also co-founding the AI Safety Club at IIT Delhi.

Has Arnav Raj published or submitted work to major ML conferences?

Yes. He submitted solo work on hallucination detection via hyperbolic geometry to ICLR 2025 Workshop. He's also a co-author on KG-MuLQA submitted to ACL 2025 (ARR) focused on multi-hop question generation and long-context evaluation.

Where is Arnav currently working?

He's currently an AI Training Data Research Intern at Abundant AI (Nov 2024 - Present), designing challenging tasks for RLHF/RL that expose failure modes in state-of-the-art LLMs. His work powers datasets used by top AI labs.

What is Arnav Raj's academic background at IIT Delhi?

Dual Degree (B.Tech + M.Tech) in Computer Science & Engineering (2022-2027) with coursework spanning ML, deep learning, information retrieval, computer architecture, operating systems, and theory of computation.

What competitive achievements does he have?

JEE Advanced AIR 1158 (top 0.1%), IMC Prosperity Trading Challenge Global Rank 8 (2025), 2× Smart India Hackathon National Top 5 Finalist, KVPY SX Fellowship, National Olympiad achievements (Top 250 Astronomy, Top 300 Chemistry), Codeforces Expert rating, and Best Mess Secretary award at IIT Delhi.