Surgan Jandial
Graduate Research Assistant
Pittsburgh, Pennsylvania, United States
Summary
👤
Senior🎓
Top SchoolSurgan Jandial is an applied ML researcher and engineer with eight years of experience building and shipping computer vision and multimodal AI systems, currently pursuing robotics research at Carnegie Mellon. At Adobe he translated research into production—deploying Deformable DETR and DinoV2 for automated forms conversion, leading LoRA fine-tuning and yielding multiple patents and top-tier publications—and he contributed core transforms and robust tests to the popular torchvision library. His recent focus blends LLM/VLM agents, multi-image reasoning, and agent safety, producing a taxonomy and synthetic evaluation framework for VLM planning that won an Amazon AWS Agentic AI award and boosting multi-image reasoning via scalable post-training pipelines. He has a track record of turning ambiguous, underdefined problems into measurable metrics and datasets (e.g., Area Under Clicks and high-accuracy synthetic GUI data) and has interned at Microsoft developing grounding-diagnosis agents for real-world desktop assistants. Based in Pittsburgh, Surgan pairs rigorous academic training from IIT Hyderabad and CMU with hands-on production experience, especially in validating model behavior and improving evaluation reliability—skills that often reveal model failure modes before they reach users.
8 years of coding experience
1 year of employment as a software developer
Bachelor of Technology - BTech, Computer Science, Bachelor of Technology - BTech, Computer Science at Indian Institute of Technology, HyderabadMaster's of Science in Robotics, 4.17/4, Master's of Science in Robotics, 4.17/4 at Carnegie Mellon University
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (8)
transform10
pytorch10
computer-vision10
python10
test-automation10
image-processing9
machine-learning8
numpy7
Programming languages (7)
Github contributions (5)
pytorch/vision
Dec 2018 - Jan 2020
Datasets, Transforms and Models specific to Computer Vision
Role in this project:
ML Engineer & Test Automation Engineer
ML Engineer & Test Automation EngineerContributions:28 commits, 53 PRs, 115 comments in 1 year 1 month
Contributions summary:Surgan primarily contributed to the `torchvision/transforms` module, focusing on computer vision transformations. They implemented new image processing modes, addressed bug fixes in the `to_tensor` function, and added functionalities, such as the `RandomPerspective` transform. Furthermore, the user significantly enhanced the testing suite by adding tests for new features and improving existing ones, including tests for out-of-place behavior and scriptability, thus demonstrating a strong focus on quality and reliability.
pytorchvisiondeep-learningdatasetcomputer-vision
surgan12/Network-hacks
May 2018 - Dec 2018
This contains all the basic networking tools , you might require .
Contributions:35 commits, 4 PRs, 35 pushes in 7 months
networknetworkingrequire
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.