CAUTION: CODE IN PRODUCTION // EXPECT HIGH EFFICIENCY
BUILDING IN THE SHADOWS // DEPLOYING TO THE LIGHT
Rohit Kumar Debnath
Software Development Engineer
DANGER: COMPILING COMPLEX ARCHITECTURES IN REAL TIME
SYSTEM OVERLOAD // ARCHITECTING FULL-STACK SOLUTIONS
Overview
SDE Intern @Hive Dynamics
Social Links
About
- Currently pursuing B.Tech in Computer Science and Engineering at Heritage Institute Of Technology with a YGPA of 9.35.
- Working as an SDE Intern at Hive Dynamics, engineering Multi-Camera Video Analytics and optimizing High-Throughput MLOps Infrastructure.
- Skilled in Python, C++, Java, JavaScript, and Solidity.
- Experienced with technologies like ReactJS, NextJS, OpenCV, Docker, NodeJS, ExpressJS, Hardhat.
- Built projects like AlgoSlice (GPU sharing on Algorand) and NanoBonds (DeFi platform for government bonds).
- Passionate about integrating AI models and streamlining high-throughput pipelines for real-world impact.
GitHub Contributions
Testimonials
Stack
Projects
Experience
Hive Dynamics
Current Employer- Engineered Multi-Camera Video Analytics: Refined YOLO and OpenCV pipelines for precise body detection and human activity tracking across active workspaces, enabling continuous automated productivity analysis.
- Implemented Robust Person ReID & Facial Recognition: Integrated ArcFace alongside deep Person Re-Identification (ReID) models to maintain consistent identity tracking of personnel across disparate camera angles and occlusions.
- Optimized High-Throughput MLOps Infrastructure: Architected a containerized Docker pipeline utilizing HLS streaming protocols and Vector Databases to ingest live camera feeds, process high-dimensional embeddings, and deliver sub-millisecond similarity matching.
- Integrated Multimodal AI Models: Implemented Vision-Language Models (VLMs) to provide semantic context to the tracked behaviors, translating raw OpenCV/YOLO coordinates into high-level, actionable workflow insights.
- Python
- OpenCV
- YOLO
- Docker
- MLOps
- Vector DB
- ArcFace
- ReID
- Computer Vision
Heritage Institute Of Technology
YGPA - 9.35 Located in Kolkata, WB, India.
Project Descriptions(2)
A prototype project on Algorand that explores remote access to GPU resources for developers who need computing power for heavy workloads. The application experiments with enabling users to utilize shared GPU infrastructure to run tasks without owning expensive hardware.
- React JS
- Python
- Rust
- CUDA
- Docker
NanoBonds is a decentralized finance (DeFi) platform that democratizes access to government bonds. It enables retail users to invest in fractional units of government bonds using stablecoins, providing a transparent, secure, and highly accessible investment experience through blockchain technology.
- ReactJS
- NodeJS
- ExpressJS
- MongoDB
- Solidity
Honors & Awards(4)
Certifications(3)
View More →
Brand
Mark
Logotype