India / IST
Portfolio / 2026
Computer Science student specialising in Artificial Intelligence and Machine Learning.
Focus
AI
Engineer
Building intelligent systems, research experiments and practical software products.
Portrait reconstructed from particles
I am Yashraj Singh, a Computer Science student specialising in Artificial Intelligence and Machine Learning. I am passionate about building intelligent systems, exploring emerging technologies, and developing software that solves real-world problems.
My interests span Artificial Intelligence, Machine Learning, Deep Learning, software engineering, and the infrastructure required to take intelligent systems into production.
I enjoy experimenting with ideas, building practical products, and continuously learning how technology can create meaningful impact.
A record of the academic stages that shaped my work in computer science, artificial intelligence and machine learning.
Assembly of Christ School, Barrackpore
Assembly of Christ School, Barrackpore
Institute of Engineering and Management, Kolkata
Bachelor of Technology/Computer Science and Engineering (Artificial Intelligence)
3rd Year
Institute of Engineering
and Management, Kolkata
B.Tech / CSE AI
India / IST
A point cloud traced from the outline of India — the place where the work, the research and the next set of ideas begin.
Scroll to explore
Vertical scroll / horizontal motion
Virtus / Scarlet
Real 0.982
Fake
0.018
Multimodal deepfake detection.
A multimodal deepfake classification framework that pairs Vision Transformers for image analysis with TimeSformer for video, exposed through a FastAPI backend and a React interface.
EizenDB
HNSW / Arweave
MCP
Memory layer / 01
Decentralised
Decentralised memory for AI agents.
A decentralised memory-sharing protocol for AI agents, pairing an Arweave-native vector database with an MCP server and a Next.js dashboard for persistent, user-owned context.
Argo / INCOIS
Ocean intelligence
SIH
18.4°N / 87.2°E
Argo ocean data, made legible.
An Argo float data visualisation platform developed for INCOIS under Smart India Hackathon, combining interactive ocean maps, parameter profiles and an AI query layer for marine data.
Flow-Anchor
Clean-label backdoor
DiT / 03
FLUX.1 / SD3
Research
Clean-label backdoors in DiTs.
A research project introducing Flow-Anchor Poisoning, a clean-label backdoor attack targeting flow-matching diffusion transformers such as FLUX.1 and Stable Diffusion 3.
Open to conversations
Have an idea, a research question or a product to build? The inbox is open.
Elsewhere
Find the code, the experiments and the work in progress.