John Zakkam
I'm a curious builder and AI enthusiast exploring the intersection of reasoning, generative models, and multimodal systems. I love tinkering with ideas that push the boundaries of what machines can understand and create. When I'm not building AI pipelines or diving into research papers, you'll find me on an exploration in the mountains or the gym.
Let's connect and build something cool together!
Work Experience
Research Engineer
OnePlus India (July 2023 - Present)
- Worked on building agentic RAG pipelines for self-testing, reducing manual workload and boosting efficiency by 15x.
- Developed transformer-based models for blind face restoration and deployed on android devices for real-time apps.
- Worked with the AI team to build a magic eraser for images, removing unwanted objects with SoTA accuracy on-device.
Education
Integrated Masters in Computer Science
IIIT Chennai (2018 - 2023)
- Worked on Deep Learning with extensive research experience in Computer Vision and GenAI
- Teching assistant for Programming in C, Pattern Recognition, Deep Learning, Image Processing and Computer Vision courses
- Final year thesis on Deepfake Detection using Contrastive Learning and Vision Transformers
Research Experience
Wichita State University
Undergraduate Research Intern (2022 - 2023)
- Collaborated with Dr. Ajita Rattani on Deepfake detection using self-supervised few-shot techniques and Vision Transformers.
- Achieved SoTA results with a Contrastive Vision Transformer model in cross-dataset deepfake detection.
- Research work CoDeiT got accepted to ICPR 2024, Kolkata.
OnePlus Research India
Summer Research Intern (2022)
- Researched GAN compression, Knowledge Distillation, Image Inpainting, and Small Object Detection.
- Mainly worked on pruning, knowledge distillation, and quantization techniques for efficient inference on mobile devices.
- Research on small object detection got accepted to ACCV 2022, Singapore.
IIITDM Kancheepuram
Computer Vision Researcher (2022)
- Developed a transformer U-Net based model for brain tumor segmentation from MR images using a curriculum learning approach.
- Worked in a funded project to detect cardiovascular diseases using Deep CNNs from retinal fundus imagery.
- Research work submitted to Elsevier, Computers in Biology and Medicine, 2024 (under review).
Publications
- S Sahayam, J Zakkam, U Jayaraman, “Can we learn better with hard samples?”, arXiv preprint arXiv:2304.03486, 2023
- J Zakkam, U Jayaraman, S Sahayam, A Rattani, “CoDeiT: Contrastive data-efficient transformers for deepfake detection”, International Conference on Pattern Recognition, 62-77, 2025 [ICPR 2024]
- A Mishra, J Zakkam, A Kumar, S Mandloi, K Anand, S Sowmya, A Thakur, “Evaluating and bench-marking object detection models for traffic sign and traffic light datasets”, Proceedings of the Asian Conference on Computer Vision, 338-353, 2022 [ACCV 2022]
- S Sahayam, J Zakkam, YS Varshan V, U Jayaraman, “Detection of Under-represented Samples Using Dynamic Batch Training for Brain Tumor Segmentation from MR Images”, arXiv preprint arXiv:2408.12013, 2024 [submitted to Computers in Biology and Medicine, Elsevier]
Technical Skills
Programming
C/C++, Python, JavaScript
Frameworks
PyTorch, Tensorflow, Keras, ONNX
Tools
VS Code, Colab, Jupyter, Git, Azure
Current Interests
LLM Research
Exploring reasoning capabilities and test-time scaling
RAG Systems
Developing efficient vector search and retrieval systems