Arkadip Maitra

GitHub Google Scholar LinkedIn X / Twitter

Reach out: arkadipmaitra at gmail dot com

Current

Working on distributed AI systems at Red Hat, contributing upstream to the PyTorch ecosystem. The work spans multi-node communication, training correctness, on-device inference, and research in computer vision, sign language, and mechanistic interpretability of reasoning in language models.

Work

Distributed training and communication form the core focus — fixing correctness and reliability issues in collective operations, transport layers, and multi-backend coordination for large-scale GPU clusters. This includes work on failure-path determinism, lifecycle management, and making communication behave predictably across hardware topologies.

On-device and edge inference work involves extending operator coverage, fixing kernel correctness, and improving build-system plumbing for selective compilation on resource-constrained targets.

Inference diagnostics and tooling contributions center on building source-level tracing and adapter infrastructure that enables any framework built on PyTorch to expose its internal call paths for debugging and profiling.

Developer tooling and framework integration work includes compiler interactions with distributed subsystems, autograd correctness for tensor subclasses, and improving error messages to surface real problems faster.

Publications

A Gloss-driven Indian Sign Language Production System using Learned Pose Representations. International Joint Conference on Artificial Intelligence (IJCAI-ECAI), 2026.

MSBAD at SignEval 2026: Multi-Scale Born-Again Distillation Ensemble for Radar-Based Sign Language Recognition. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2026.

RAVEN: A Rapid Agentic Vision Framework for Emergency Response in Vulnerable Settlements. IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, 2026.

Enhancing Earthquake Preparedness in the Himalayan Region: A Machine Learning Approach using EEW System Parameters. Iranian Journal of Science and Technology, Transactions of Electrical Engineering, 2026.

Hierarchical Windowed Graph Attention Transformer Encoder and a Large Scale Dataset for Indian Sign Language Recognition. Pattern Analysis and Applications, 2025.

Decorrelation-Based Self-Supervised Visual Representation Learning for Writer Identification. ACM Transactions on Asian and Low-Resource Language Information Processing, 2025.

A Continuous Indian Sign Language Production System from Isolated Signs. International Conference on Smart Computing and Communications (ICSCC), 2025.

Edge-Based Intelligent and Smart Health Monitoring on PYNQ-Z2 using Lightweight Protocol and Integration of Machine Learning Models. International Journal on Smart Sensing and Intelligent Systems, 2025.

Community

Served as a reviewer for IEEE Transactions on AI, the official PyTorch blog, and the PyTorch Associate certification course. Program committee member for MCP Developer Conf Bangalore and CFP reviewer for MCP Dev Summit across Mumbai, Bangalore, and Seoul.

Gave talks on distributed communication, PyTorch internals, and collective operations at BangPypers, Razorpay Engineering Meetup, HuggingFace Meetup Bangalore, and multiple university sessions. Super mentor at the HuggingFace-Meta Hackathon.

Published author on the Red Hat Developer blog, writing on distributed communication infrastructure in PyTorch.