Scholar
Swaroop Mishra
Google Scholar ID: -7LK2SwAAAAJ
Research Scientist, Google DeepMind
Citations & Impact
Citations
22,566
H-index
34
i10-index
55
Publications
20
Co-authors
7
list available
Publications
Resume (English only)
Academic Achievements
- Pioneered instruction-tuning with the first published paper on the topic.
- Key contributions include: Super-NaturalInstructions (scaling), Biomedical Instruction-tuning (applications), Instruction-following Evals (evaluation), and instruction-bias (bias analysis, Outstanding Paper Award at EACL 2023).
- Developed reasoning methods such as reframing, question-decomposition, math via programs, help me think, and take a step back; created datasets including NumGLUE, FeasibilityQA, and TarGEN.
- Proposed Self-Instruct (training) and Self-Discover (evaluation) for model self-improvement, and analyzed limitations of self-correction.
- Awards: GDM Impact Award 2025 (for IMO 2024 Silver Medal contribution), Odia Bhasa Samman 2025, ASU Dean’s Dissertation Award 2023, CEN-CS Outstanding Student Award 2023.
- Paper "Natural Instructions" received the AI2 "Lasting Impact Paper Award" 3.5 years after publication.
- Co-organized the MATH-AI workshop at NeurIPS 2022 and NeurIPS 2024.
- Holds multiple patents on LLM reasoning, scaling, post-training, self-improvement, and inference.
Miscellany
- Featured in media coverage including The New York Times (IMO Silver Medal), Kanak News interview, and profile "The Making of a Scientist: From Classroom Rebel to AI Innovator".
- Participated in podcasts on research initiation, instruction tuning, prompt engineering, and self-improving LLMs, including the NIST Podcast Series.
- Engaged in panel discussions at NeurIPS and ACL with leading AI researchers; delivered invited talks at Allen AI, IIT Kanpur, University of Toronto, Google Research, and Facebook AI Research.
- Interviewed Jeff Dean at the Odisha AI Conference 2024.
- Personal journey highlighted in the "Ones Changing The World - 1CW" podcast.
