CV
Basics
| Name | Gunjan Dhanuka |
| Label | AI Research Scientist |
| gdhanuka192@gmail.com | |
| Url | https://gunjandhanuka.github.io |
| Summary | AI Research Scientist at Mistral AI, working on multimodal models for document understanding, computer use, and everyday tasks. |
Work
-
2026.02 - Present Palo Alto, CA
AI Research Scientist
Mistral AI
Researching multimodal models for document understanding and computer use.
- Core team member behind Mistral OCR 4 and OCR 4.1, state-of-the-art OCR models for document intelligence.
- Working on OCR and document-understanding models.
- Working on computer-use capabilities.
-
2025.05 - 2025.08 Cupertino, CA
Machine Learning Engineer Intern
Apple
Worked with multimodal LLM agents for automated data analysis.
-
2025.01 - 2025.12 Pittsburgh, PA
Graduate Teaching Assistant
Machine Learning Department, Carnegie Mellon University
Teaching assistant for ML with Large Datasets.
-
2025.01 - 2026.01 Pittsburgh, PA
Graduate Student Researcher
Machine Learning Department, Carnegie Mellon University
Advised by Zico Kolter and Pratyush Maini on compression-based methods for LLMs and memorization in diffusion models.
- Collaborated with OpenAI on compression-based methods for LLMs.
- Proposed a compression-based definition of memorization in diffusion models; accepted to the ICML MemFM Workshop 2025 and submitted to NeurIPS 2025.
-
2024.12 - 2025.05 Pittsburgh, PA
Graduate Student Researcher
Robotics Institute, Carnegie Mellon University
Researched unsupervised, customizable model debiasing using meta-learning with Fernando de la Torre and Shayok Chakraborty.
- Submitted the work to NeurIPS 2025.
-
2023.05 - 2023.07 Bengaluru, India
Software Engineer Intern
Rubrik
Worked on horizontal scaling for on-premises systems.
- Devised a leader-worker architecture that improved core-service performance by 5x.
- Developed a Go and Kubernetes CLI for adding and removing cluster nodes without downtime.
-
2022.05 - 2024.01 Pittsburgh, PA
Research Intern
Carnegie Mellon University
Worked on weakly supervised video anomaly detection.
- Outperformed prior results on XD-Violence, UCF-Crime, and ShanghaiTech; accepted to WACV 2025.
-
2021.11 - 2022.05 Remote
Research Intern
UNSW
Modelled SARS-CoV-2 spike-protein mutation tendencies with graph-theoretic methods in collaboration with CSIRO Australia.
Education
-
2024.08 - 2025.12 Pittsburgh, PA, USA
Master of Science
Carnegie Mellon University
Computer Science
- Machine Learning
- ML Systems
- Distributed Systems
-
2020.11 - 2024.05 Assam, India
Bachelor of Technology
Indian Institute of Technology, Guwahati
Computer Science and Engineering; minor in Robotics and Artificial Intelligence
- Computer Systems
- Machine Learning
- Parallel Algorithms
- Computer Networks