Mir Nafis Sharear Shopnil

Mir Nafis Sharear Shopnil

AI Engineer & Research Fellow

Technovative Solutions Ltd • Fatima Fellowship

About

I am an AI researcher bridging the gap between Multi-Agent Systems, Large Language Models, and real-world reliability.

Currently, I work as an AI Engineer at Technovative Solutions, architecting agentic workflows for climate adaptation, and as a Research Fellow at Fatima Fellowship, developing verifiable misinformation detection pipelines.

Python PyTorch LangChain GNNs FastAPI AWS

Research Focus

🤖 Agentic AI

Orchestrating LLMs to solve complex, multi-step reasoning tasks with tool use.

🕵️ Misinformation

Building multimodal verification pipelines using verifiable evidence graphs.

🕸 Graph Learning

Utilizing GCNs/GATs for structured reasoning and link prediction.

🧠 Interpretability

Demystifying model decisions through mechanistic analysis and rationales.

Selected Works

MIRAGE: Agentic Framework for Misinformation Detection

Under Review @ ACM WebConf 2026

A pipeline orchestrating LVLMs with web-search to build verifiable evidence graphs.

Evidence-Grounded Multimodal Detection with GNNs

Under Review @ ACL ARR 2025

Cross-graph attention networks for image-text consistency scoring. (93.05% Accuracy)

Updates

  • Oct 2025 Submitted MIRAGE framework to ACM WebConf.
  • Aug 2025 Joined Technovative Solutions as AI Engineer (Lead Architect).
  • Sep 2024 Selected as Research Fellow at Fatima Fellowship.